diff --git a/WINNING_RUNBOOK.md b/WINNING_RUNBOOK.md new file mode 100644 index 0000000000..36b0181b2e --- /dev/null +++ b/WINNING_RUNBOOK.md @@ -0,0 +1,89 @@ +# Parameter Golf Winning Runbook ($25 Edition) + +This runbook is optimized for low budget and high decision quality. +Goal: make one real SOTA attempt without wasting credits. + +## 0) Non-negotiables + +- Beat current SOTA with margin and significance, not one lucky seed. +- Keep artifact under `16,000,000` bytes (decimal MB). +- Never use validation or train data illegally during quantization/eval. +- Prefer cheap filtering first, then expensive confirmation. + +## 1) Budget split + +- Phase A (cheap filtering): `$8` +- Phase B (1xH100 confirmation): `$9` +- Phase C (final 8xH100 reproducibility): `$8` (or wait for grant) + +If you get OpenAI credits, expand Phase C to 3-seed evidence. + +## 2) Exact baseline to start from + +Use this folder as your starting point: + +- `records/track_10min_16mb/2026-03-25_ValCalib_GPTQ_XSA_BigramHash3072/` + +Do not start from old baseline scripts. + +## 3) Runpod setup (first pod) + +1. Launch cheap single-GPU pod first (L40/4090/5090 class). +2. SSH in and run: + - `cd /workspace` + - `git clone https://github.com/openai/parameter-golf.git` + - `cd parameter-golf` + - `python3 data/cached_challenge_fineweb.py --variant sp1024 --train-shards 1` +3. Copy your working `train_gpt.py` candidate into a new local work folder. + +## 4) Experiment matrix (run in this order) + +Use `commands/runpod_experiments.sh`. + +Design principle: +- Change one high-impact axis at a time. +- Keep all other vars fixed. +- Promote only stable gains. + +Priority axes: +- `GPTQ_CALIB_BATCHES`: 192/256/320 +- `GPTQ_BLOCK_SIZE`: 128/256 +- `BIGRAM_DIM`: 96/112/128 (with `BIGRAM_VOCAB_SIZE=3072`) +- `WARMDOWN_ITERS`: 3500/4000/4500 +- `TARGET_MB`: 15.85/15.90 + +## 5) Stop/go rules (strict) + +- If run regresses by `>= 0.0015 bpb` vs control: stop that branch. +- If run improves by `< 0.0007 bpb`: do not promote. +- Promote only if improved in 2 seeds (cheap pod is fine for this check). +- Spend H100 only on top 1-2 configs. + +## 6) Promotion checklist before H100 + +- Script runs clean with no dependency errors. +- Final lines print `val_bpb` and compressed model size. +- Artifact clearly below 16MB target. +- No rule-violating data access in quantization/eval path. + +## 7) Submission checklist + +Create a new folder in `records/track_10min_16mb/_/` with: + +- `README.md` (what changed, why, exact command) +- `submission.json` +- `train_gpt.py` +- `train.log` (or multiple logs for significance) +- `requirements.txt` only if non-default deps were needed + +## 8) Your daily cadence (copy exactly) + +1. 6 cheap ablations (Phase A). +2. Pick top 2 and re-run with new seeds. +3. Move best to 1xH100 (Phase B). +4. If still positive, run final reproducibility pass (Phase C or grant credits). +5. Submit PR the same day while evidence is fresh. + +## 9) One hard rule + +Do not chase tiny LR/WD decimal tweaks until your quantization + calibration stack is already clearly beating your current best. diff --git a/commands/run_remaining_after_a2_4090.sh b/commands/run_remaining_after_a2_4090.sh new file mode 100644 index 0000000000..9b09ecbf2e --- /dev/null +++ b/commands/run_remaining_after_a2_4090.sh @@ -0,0 +1,53 @@ +#!/usr/bin/env bash +# Run experiments b1, c1, c2, d1, d2 after ctrl + a1 + a2 are done (1x RTX 4090, low VRAM). +# Usage: bash commands/run_remaining_after_a2_4090.sh /path/to/train_gpt.py +set -euo pipefail + +TRAIN_SCRIPT="${1:-records/track_10min_16mb/2026-03-25_ValCalib_GPTQ_XSA_BigramHash3072/train_gpt.py}" + +if [[ ! -f "${TRAIN_SCRIPT}" ]]; then + echo "ERROR: train script not found: ${TRAIN_SCRIPT}" + exit 1 +fi + +export TRAIN_BATCH_TOKENS="${TRAIN_BATCH_TOKENS:-196608}" +export PYTORCH_CUDA_ALLOC_CONF="${PYTORCH_CUDA_ALLOC_CONF:-expandable_segments:True}" +export DATA_PATH="${DATA_PATH:-./data/datasets/fineweb10B_sp1024/}" +export TOKENIZER_PATH="${TOKENIZER_PATH:-./data/tokenizers/fineweb_1024_bpe.model}" +export VOCAB_SIZE="${VOCAB_SIZE:-1024}" +export MAX_WALLCLOCK_SECONDS="${MAX_WALLCLOCK_SECONDS:-600}" +export TRAIN_SEQ_LEN="${TRAIN_SEQ_LEN:-2048}" +export EVAL_SEQ_LEN="${EVAL_SEQ_LEN:-2048}" +export VAL_LOSS_EVERY="${VAL_LOSS_EVERY:-0}" +export TARGET_MB="${TARGET_MB:-15.9}" +export BIGRAM_VOCAB_SIZE="${BIGRAM_VOCAB_SIZE:-3072}" +export XSA_LAST_N="${XSA_LAST_N:-11}" + +mkdir -p logs + +run_one() { + local run_id="$1" + shift + echo "==== START ${run_id} ====" + RUN_ID="${run_id}" "$@" torchrun --standalone --nproc_per_node=1 "${TRAIN_SCRIPT}" \ + 2>&1 | tee "logs/${run_id}.log" + echo "==== END ${run_id} ====" +} + +run_one "b1_block256_seed314" env \ + SEED=314 WARMDOWN_ITERS=4000 GPTQ_CALIB_BATCHES=256 GPTQ_BLOCK_SIZE=256 BIGRAM_DIM=112 + +run_one "c1_bigram96_seed314" env \ + SEED=314 WARMDOWN_ITERS=4000 GPTQ_CALIB_BATCHES=256 GPTQ_BLOCK_SIZE=128 BIGRAM_DIM=96 TARGET_MB=15.85 + +run_one "c2_bigram128_seed314" env \ + SEED=314 WARMDOWN_ITERS=4000 GPTQ_CALIB_BATCHES=256 GPTQ_BLOCK_SIZE=128 BIGRAM_DIM=128 TARGET_MB=15.90 + +run_one "d1_warm3500_seed314" env \ + SEED=314 WARMDOWN_ITERS=3500 GPTQ_CALIB_BATCHES=256 GPTQ_BLOCK_SIZE=128 BIGRAM_DIM=112 + +run_one "d2_warm4500_seed314" env \ + SEED=314 WARMDOWN_ITERS=4500 GPTQ_CALIB_BATCHES=256 GPTQ_BLOCK_SIZE=128 BIGRAM_DIM=112 + +echo "All five runs done." +python3 commands/summarize_logs.py logs diff --git a/commands/runpod_experiments.sh b/commands/runpod_experiments.sh new file mode 100644 index 0000000000..7c27c700cf --- /dev/null +++ b/commands/runpod_experiments.sh @@ -0,0 +1,73 @@ +#!/usr/bin/env bash +set -euo pipefail + +# Usage: +# bash commands/runpod_experiments.sh /path/to/train_gpt.py +# +# Example: +# bash commands/runpod_experiments.sh records/track_10min_16mb/2026-03-25_ValCalib_GPTQ_XSA_BigramHash3072/train_gpt.py +# +# This script runs a low-cost ablation matrix on 1 GPU. + +TRAIN_SCRIPT="${1:-train_gpt.py}" + +if [[ ! -f "${TRAIN_SCRIPT}" ]]; then + echo "ERROR: train script not found: ${TRAIN_SCRIPT}" + exit 1 +fi + +export DATA_PATH="${DATA_PATH:-./data/datasets/fineweb10B_sp1024/}" +export TOKENIZER_PATH="${TOKENIZER_PATH:-./data/tokenizers/fineweb_1024_bpe.model}" +export VOCAB_SIZE="${VOCAB_SIZE:-1024}" +export MAX_WALLCLOCK_SECONDS="${MAX_WALLCLOCK_SECONDS:-600}" +export TRAIN_BATCH_TOKENS="${TRAIN_BATCH_TOKENS:-786432}" +export TRAIN_SEQ_LEN="${TRAIN_SEQ_LEN:-2048}" +export EVAL_SEQ_LEN="${EVAL_SEQ_LEN:-2048}" +export VAL_LOSS_EVERY="${VAL_LOSS_EVERY:-0}" +export TARGET_MB="${TARGET_MB:-15.9}" +export BIGRAM_VOCAB_SIZE="${BIGRAM_VOCAB_SIZE:-3072}" +export XSA_LAST_N="${XSA_LAST_N:-11}" + +mkdir -p logs + +run_one() { + local run_id="$1" + shift + echo "==== START ${run_id} ====" + RUN_ID="${run_id}" "$@" torchrun --standalone --nproc_per_node=1 "${TRAIN_SCRIPT}" \ + 2>&1 | tee "logs/${run_id}.log" + echo "==== END ${run_id} ====" +} + +# Control +run_one "ctrl_seed314" env \ + SEED=314 WARMDOWN_ITERS=4000 GPTQ_CALIB_BATCHES=256 GPTQ_BLOCK_SIZE=128 BIGRAM_DIM=112 + +# A1/A2: calibration coverage +run_one "a1_calib192_seed314" env \ + SEED=314 WARMDOWN_ITERS=4000 GPTQ_CALIB_BATCHES=192 GPTQ_BLOCK_SIZE=128 BIGRAM_DIM=112 + +run_one "a2_calib320_seed314" env \ + SEED=314 WARMDOWN_ITERS=4000 GPTQ_CALIB_BATCHES=320 GPTQ_BLOCK_SIZE=128 BIGRAM_DIM=112 + +# B1: GPTQ block size +run_one "b1_block256_seed314" env \ + SEED=314 WARMDOWN_ITERS=4000 GPTQ_CALIB_BATCHES=256 GPTQ_BLOCK_SIZE=256 BIGRAM_DIM=112 + +# C1/C2: bigram dim tradeoff +run_one "c1_bigram96_seed314" env \ + SEED=314 WARMDOWN_ITERS=4000 GPTQ_CALIB_BATCHES=256 GPTQ_BLOCK_SIZE=128 BIGRAM_DIM=96 TARGET_MB=15.85 + +run_one "c2_bigram128_seed314" env \ + SEED=314 WARMDOWN_ITERS=4000 GPTQ_CALIB_BATCHES=256 GPTQ_BLOCK_SIZE=128 BIGRAM_DIM=128 TARGET_MB=15.90 + +# D1/D2: warmdown schedule +run_one "d1_warm3500_seed314" env \ + SEED=314 WARMDOWN_ITERS=3500 GPTQ_CALIB_BATCHES=256 GPTQ_BLOCK_SIZE=128 BIGRAM_DIM=112 + +run_one "d2_warm4500_seed314" env \ + SEED=314 WARMDOWN_ITERS=4500 GPTQ_CALIB_BATCHES=256 GPTQ_BLOCK_SIZE=128 BIGRAM_DIM=112 + +echo "All runs done. Logs are in logs/." +echo "Next: parse results quickly with:" +echo " rg -n \"val_bpb|final_int8_zlib_roundtrip|artifact|compressed\" logs/*.log" diff --git a/commands/summarize_logs.py b/commands/summarize_logs.py new file mode 100644 index 0000000000..6d2cf8808d --- /dev/null +++ b/commands/summarize_logs.py @@ -0,0 +1,59 @@ +#!/usr/bin/env python3 +""" +Summarize Parameter Golf train logs. Prefer leaderboard-style sliding BPB +(final_int6_sliding_window_exact) over plain val_bpb or roundtrip lines. +""" +import re +import sys +from pathlib import Path + +# Leaderboard-style metric (see record README "Sliding BPB") +SLIDING_EXACT = re.compile( + r"final_int6_sliding_window_exact val_loss:[0-9.e+-]+ val_bpb:([0-9.]+)" +) +SLIDING_S64_EXACT = re.compile( + r"final_int6_sliding_window_s64_exact val_loss:[0-9.e+-]+ val_bpb:([0-9.]+)" +) +ROUNDTRIP_EXACT = re.compile( + r"final_int6_roundtrip_exact val_loss:[0-9.e+-]+ val_bpb:([0-9.]+)" +) +VAL_ANY = re.compile(r"val_bpb:([0-9.]+)") +ARTIFACT_RE = re.compile( + r"Total submission size int6\+lzma:\s*([0-9]+)\s*bytes?", re.IGNORECASE +) + + +def parse_file(path: Path): + text = path.read_text(encoding="utf-8", errors="ignore") + + val = "NA" + for pattern in (SLIDING_EXACT, SLIDING_S64_EXACT, ROUNDTRIP_EXACT): + m = pattern.search(text) + if m: + val = m.group(1) + break + if val == "NA": + vals = VAL_ANY.findall(text) + val = vals[-1] if vals else "NA" + + art_m = ARTIFACT_RE.search(text) + art = art_m.group(1) if art_m else "NA" + + return val, art + + +def main(): + log_dir = Path(sys.argv[1]) if len(sys.argv) > 1 else Path("logs") + files = sorted(log_dir.glob("*.log")) + if not files: + print(f"No log files found in {log_dir}") + return + + print("run_id,sliding_val_bpb_or_best_available,artifact_bytes_int6_lzma") + for f in files: + val, art = parse_file(f) + print(f"{f.stem},{val},{art}") + + +if __name__ == "__main__": + main() diff --git a/records/track_10min_16mb/2026-03-20_10L_Int5MLP_MuonWD04_SWA50/README.md b/records/track_10min_16mb/2026-03-20_10L_Int5MLP_MuonWD04_SWA50/README.md index d06e11fc62..041e3c87ea 100644 --- a/records/track_10min_16mb/2026-03-20_10L_Int5MLP_MuonWD04_SWA50/README.md +++ b/records/track_10min_16mb/2026-03-20_10L_Int5MLP_MuonWD04_SWA50/README.md @@ -5,7 +5,7 @@ ## Run Command ```bash -# Setup (once) +# Setup (once): downloads tokenizer + training/val shards (default 80 train shards) bash prepare.sh # Train + evaluate (default seed=42) @@ -13,9 +13,14 @@ bash eval/eval.sh # With specific seed SEED=42 bash eval/eval.sh + +# Quick smoke test (few steps, small batch; set SWA_ENABLED=0) — not a leaderboard score +bash eval/smoke.sh ``` -All parameters are set as defaults in `train_gpt.py`. No env vars needed. +All parameters are set as defaults in `train_gpt.py`. No env vars needed for a full run. + +For a smaller download while iterating locally, run `TRAIN_SHARDS=1 bash prepare.sh` from this directory (or pass `--train-shards 1` to `data/cached_challenge_fineweb.py` from the repo root). ## 3-Seed Results diff --git a/records/track_10min_16mb/2026-03-20_10L_Int5MLP_MuonWD04_SWA50/eval/eval.sh b/records/track_10min_16mb/2026-03-20_10L_Int5MLP_MuonWD04_SWA50/eval/eval.sh new file mode 100644 index 0000000000..e0b4e1ff8f --- /dev/null +++ b/records/track_10min_16mb/2026-03-20_10L_Int5MLP_MuonWD04_SWA50/eval/eval.sh @@ -0,0 +1,13 @@ +#!/usr/bin/env bash +# Full leaderboard training run (defaults match submission; ~10 min cap on 1xH100). +set -euo pipefail +RECORD_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)" +ROOT="$(cd "$RECORD_DIR/../../.." && pwd)" +cd "$ROOT" +export DATA_PATH="${DATA_PATH:-$ROOT/data/datasets/fineweb10B_sp1024}" +export TOKENIZER_PATH="${TOKENIZER_PATH:-$ROOT/data/tokenizers/fineweb_1024_bpe.model}" +export VOCAB_SIZE="${VOCAB_SIZE:-1024}" +export RUN_ID="${RUN_ID:-10L_Int5MLP_MuonWD04_SWA50}" +export SEED="${SEED:-42}" +NPROC="${NPROC:-1}" +exec torchrun --standalone --nproc_per_node="${NPROC}" "${RECORD_DIR}/train_gpt.py" diff --git a/records/track_10min_16mb/2026-03-20_10L_Int5MLP_MuonWD04_SWA50/eval/smoke.sh b/records/track_10min_16mb/2026-03-20_10L_Int5MLP_MuonWD04_SWA50/eval/smoke.sh new file mode 100644 index 0000000000..8117446044 --- /dev/null +++ b/records/track_10min_16mb/2026-03-20_10L_Int5MLP_MuonWD04_SWA50/eval/smoke.sh @@ -0,0 +1,21 @@ +#!/usr/bin/env bash +# Tiny run to verify CUDA, data paths, and script wiring (not a score attempt). +set -euo pipefail +RECORD_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)" +ROOT="$(cd "$RECORD_DIR/../../.." && pwd)" +cd "$ROOT" +export DATA_PATH="${DATA_PATH:-$ROOT/data/datasets/fineweb10B_sp1024}" +export TOKENIZER_PATH="${TOKENIZER_PATH:-$ROOT/data/tokenizers/fineweb_1024_bpe.model}" +export VOCAB_SIZE="${VOCAB_SIZE:-1024}" +export RUN_ID="${RUN_ID:-smoke_10L_int5}" +export SEED="${SEED:-42}" +export ITERATIONS="${ITERATIONS:-8}" +export WARMUP_STEPS="${WARMUP_STEPS:-2}" +export WARMDOWN_ITERS="${WARMDOWN_ITERS:-4}" +export TRAIN_BATCH_TOKENS="${TRAIN_BATCH_TOKENS:-65536}" +export MAX_WALLCLOCK_SECONDS="${MAX_WALLCLOCK_SECONDS:-0}" +export VAL_LOSS_EVERY="${VAL_LOSS_EVERY:-0}" +export TRAIN_LOG_EVERY="${TRAIN_LOG_EVERY:-1}" +export SWA_ENABLED="${SWA_ENABLED:-0}" +NPROC="${NPROC:-1}" +exec torchrun --standalone --nproc_per_node="${NPROC}" "${RECORD_DIR}/train_gpt.py" diff --git a/records/track_10min_16mb/2026-03-20_10L_Int5MLP_MuonWD04_SWA50/prepare.sh b/records/track_10min_16mb/2026-03-20_10L_Int5MLP_MuonWD04_SWA50/prepare.sh new file mode 100644 index 0000000000..ee8ee0daf3 --- /dev/null +++ b/records/track_10min_16mb/2026-03-20_10L_Int5MLP_MuonWD04_SWA50/prepare.sh @@ -0,0 +1,7 @@ +#!/usr/bin/env bash +# Download FineWeb shards + tokenizer into ./data/ (run from anywhere). +set -euo pipefail +ROOT="$(cd "$(dirname "${BASH_SOURCE[0]}")/../../.." && pwd)" +cd "$ROOT" +TRAIN_SHARDS="${TRAIN_SHARDS:-80}" +exec python3 data/cached_challenge_fineweb.py --variant sp1024 --train-shards "${TRAIN_SHARDS}" diff --git a/records/track_10min_16mb/2026-03-20_10L_Int5MLP_MuonWD04_SWA50/train_gpt.py b/records/track_10min_16mb/2026-03-20_10L_Int5MLP_MuonWD04_SWA50/train_gpt.py index bbe5ab2943..01c6e8e829 100644 --- a/records/track_10min_16mb/2026-03-20_10L_Int5MLP_MuonWD04_SWA50/train_gpt.py +++ b/records/track_10min_16mb/2026-03-20_10L_Int5MLP_MuonWD04_SWA50/train_gpt.py @@ -851,8 +851,10 @@ def main() -> None: from torch.backends.cuda import enable_cudnn_sdp, enable_flash_sdp, enable_math_sdp, enable_mem_efficient_sdp enable_cudnn_sdp(False) enable_flash_sdp(True) - enable_mem_efficient_sdp(False) - enable_math_sdp(False) + # Flash is unavailable on many Windows CUDA builds; without math/mem SDPA errors ("No available kernel"). + sdp_fallback = bool(int(os.environ.get("ENABLE_MATH_SDP", "0"))) or sys.platform == "win32" + enable_mem_efficient_sdp(sdp_fallback) + enable_math_sdp(sdp_fallback) logfile = None if master_process: diff --git a/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/README.md b/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/README.md new file mode 100644 index 0000000000..2f5f8e6dc8 --- /dev/null +++ b/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/README.md @@ -0,0 +1,64 @@ +# Non-record: 4090 ablations on ValCalib GPTQ + XSA + BigramHash stack + +**Author:** Wolfie8935 +**Intent:** Document a **budget hardware** exploration (single **RTX 4090**, 24GB VRAM) on the same code lineage as the 10-minute leaderboard stack ([`2026-03-25_ValCalib_GPTQ_XSA_BigramHash3072`](../track_10min_16mb/2026-03-25_ValCalib_GPTQ_XSA_BigramHash3072/README.md)), **without** claiming a new 8×H100 record. + +This is **not** comparable to the public SOTA (~**1.1147** sliding BPB on 8×H100 SXM): we used **`TRAIN_BATCH_TOKENS=196608`** (OOM-safe on 24GB) instead of the multi-GPU throughput recipe, so step counts and trajectories differ. + +## What is in this folder + +| File | Purpose | +|------|---------| +| `train_gpt.py` | Unmodified snapshot from the ValCalib / PR #1019 lineage (same as reference record). | +| `requirements.txt` | Same as reference record (FlashAttention 3 install: see below). | +| `train_*.log` | Raw Runpod logs for completed ablations (see caveats). | + +## Ablations attempted (seed 314, 600s wallclock) + +| Run ID | Knobs | Notes | +|--------|--------|--------| +| `ctrl` | Default stack, `GPTQ_CALIB_BATCHES=256`, `BLOCK=128`, `BIGRAM_DIM=112`, `WARMDOWN=4000` | Log ends with **SIGTERM** after `final_int6_roundtrip_exact` — **sliding eval did not complete** in this capture. | +| `a1` | `GPTQ_CALIB_BATCHES=192` | **Has** `final_int6_sliding_window_exact` → use for best **reported** sliding BPB in this bundle. | +| `a2` | `GPTQ_CALIB_BATCHES=320` | Completed through roundtrip; verify log for sliding line when using for analysis. | +| `b1` | `GPTQ_BLOCK_SIZE=256` | Completed through roundtrip in log snapshot. | + +**Not included (incomplete or not finished at time of packaging):** `c1`, `c2`, `d1`, `d2` — you can add `train_*.log` files later and update `submission.json`. + +## Best reported metric in included logs (for transparency) + +From **`train_a1_calib192_seed314.log`**: + +- `final_int6_sliding_window_exact` → **val_bpb ≈ 1.56854187** (post-quant sliding window; see log line). + +This number is **much worse** than leaderboard SOTA; it reflects **4090 + smaller microbatch**, not a failure of the upstream architecture. + +## Reproduce (single GPU, Runpod-style) + +```bash +export TRAIN_BATCH_TOKENS=196608 +export PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True +export DATA_PATH=./data/datasets/fineweb10B_sp1024/ +export TOKENIZER_PATH=./data/tokenizers/fineweb_1024_bpe.model +export VOCAB_SIZE=1024 +export MAX_WALLCLOCK_SECONDS=600 +export BIGRAM_VOCAB_SIZE=3072 +export XSA_LAST_N=11 + +# Example: control config +RUN_ID=ctrl_seed314 SEED=314 WARMDOWN_ITERS=4000 GPTQ_CALIB_BATCHES=256 GPTQ_BLOCK_SIZE=128 BIGRAM_DIM=112 \ +torchrun --standalone --nproc_per_node=1 train_gpt.py +``` + +**FlashAttention 3 (Hopper-focused in upstream README):** on 4090 you may need the image’s installed FA / PyTorch stack; if imports fail, install per the reference record’s `README.md`. + +## Why non-record + +- Does not meet **8×H100 / 10-minute official** leaderboard bar. +- Incomplete ablation matrix and incomplete **`ctrl` log** (no sliding line). +- Honest documentation of **single-GPU** constraints. + +## Next steps (if pursuing a real record later) + +1. Re-run **`ctrl`** on 8×H100 without interrupting the process until **`final_int6_sliding_window_exact`** prints. +2. Complete **`c1`–`d2`** or drop them from the story. +3. Run **3 seeds** and collect significance vs prior SOTA per challenge rules. diff --git a/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/requirements.txt b/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/requirements.txt new file mode 100644 index 0000000000..8b0f870b9b --- /dev/null +++ b/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/requirements.txt @@ -0,0 +1,3 @@ +# FlashAttention 3 must be installed separately; see README.md +sentencepiece +zstandard diff --git a/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/submission.json b/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/submission.json new file mode 100644 index 0000000000..5da1d278d9 --- /dev/null +++ b/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/submission.json @@ -0,0 +1,29 @@ +{ + "author": "Wolfie8935", + "github_id": "Wolfie8935", + "name": "4090 ablations on ValCalib GPTQ+XSA stack (incomplete matrix)", + "blurb": "Non-record submission documenting single RTX 4090 ablations (TRAIN_BATCH_TOKENS=196608) on the 2026-03-25 ValCalib / AR self-gen GPTQ + XSA-all + BigramHash 3072x112 lineage. Includes logs for ctrl, a1, a2, b1 only; c1/c2/d1/d2 not packaged. Best included sliding BPB from train_a1_calib192_seed314.log: final_int6_sliding_window_exact val_bpb ~1.56854187 — not comparable to 8xH100 SOTA. Ctrl log interrupted before sliding eval in captured run.", + "date": "2026-04-01", + "track": "non-record-16mb", + "submission_kind": "non_record", + "record_claim": false, + "val_loss": null, + "val_bpb": 1.56854187, + "val_bpb_metric": "final_int6_sliding_window_exact (from train_a1_calib192_seed314.log only)", + "val_bpb_note": "4090 single-GPU reduced microbatch; not comparable to 8xH100 leaderboard", + "seeds": [314], + "hardware": "1x NVIDIA GeForce RTX 4090 24GB", + "wallclock_seconds": 600, + "train_batch_tokens": 196608, + "lineage": "records/track_10min_16mb/2026-03-25_ValCalib_GPTQ_XSA_BigramHash3072", + "included_logs": [ + "train_ctrl_seed314.log", + "train_a1_calib192_seed314.log", + "train_a2_calib320_seed314.log", + "train_b1_block256_seed314.log" + ], + "missing_ablations": ["c1_bigram96_seed314", "c2_bigram128_seed314", "d1_warm3500_seed314", "d2_warm4500_seed314"], + "known_log_issues": [ + "train_ctrl_seed314.log: process received SIGTERM before final_int6_sliding_window_exact in captured log" + ] +} diff --git a/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/train_a1_calib192_seed314.log b/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/train_a1_calib192_seed314.log new file mode 100644 index 0000000000..e7a151374a --- /dev/null +++ b/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/train_a1_calib192_seed314.log @@ -0,0 +1,70 @@ +logs/a1_calib192_seed314.txt +val_bpb:enabled tokenizer_kind=sentencepiece tokenizer_path=./data/tokenizers/fineweb_1024_bpe.model +train_loader:dataset:fineweb10B_sp1024 train_shards:1 +val_loader:shards pattern=./data/datasets/fineweb10B_sp1024/fineweb_val_*.bin tokens:62021632 +model_params:27067484 +mtp_num_heads:0 mtp_loss_weight:0.2 mtp_params:0 +XSA:last_11 active_layers:[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10] +world_size:1 grad_accum_steps:8 +sdp_backends:cudnn=False flash=True mem_efficient=False math=False +attention_mode:gqa num_heads:8 num_kv_heads:4 +tie_embeddings:True embed_lr:0.035 head_lr:0.0 matrix_lr:0.025 scalar_lr:0.025 +train_batch_tokens:196608 train_seq_len:2048 iterations:20000 warmup_steps:20 max_wallclock_seconds:600.000 +seed:314 +warmup_step:1/20 +warmup_step:2/20 +warmup_step:3/20 +warmup_step:4/20 +warmup_step:5/20 +warmup_step:6/20 +warmup_step:7/20 +warmup_step:8/20 +warmup_step:9/20 +warmup_step:10/20 +warmup_step:11/20 +warmup_step:12/20 +warmup_step:13/20 +warmup_step:14/20 +warmup_step:15/20 +warmup_step:16/20 +warmup_step:17/20 +warmup_step:18/20 +warmup_step:19/20 +warmup_step:20/20 +step:1/20000 train_loss:6.9300 train_time:528ms step_avg:528.48ms +step:2/20000 train_loss:8.7825 train_time:1001ms step_avg:500.65ms +step:3/20000 train_loss:8.3371 train_time:1517ms step_avg:505.63ms +step:4/20000 train_loss:7.6582 train_time:2035ms step_avg:508.75ms +step:5/20000 train_loss:7.0457 train_time:2553ms step_avg:510.67ms +step:6/20000 train_loss:6.5797 train_time:3083ms step_avg:513.81ms +step:7/20000 train_loss:6.3267 train_time:3602ms step_avg:514.51ms +step:8/20000 train_loss:6.0421 train_time:4120ms step_avg:515.04ms +step:9/20000 train_loss:5.9837 train_time:5124ms step_avg:569.34ms +step:10/20000 train_loss:5.9126 train_time:5644ms step_avg:564.39ms +swa:start step:400 +step:500/20000 train_loss:2.5951 train_time:256286ms step_avg:512.57ms +late_qat:enabled step:567 scale:0.1500 +step:1000/20000 train_loss:2.3988 train_time:513261ms step_avg:513.26ms +step:1168/20000 val_loss:2.3636 val_bpb:1.3999 train_time:600409ms step_avg:514.05ms +stopping_early: wallclock_cap train_time:600409ms step:1168/20000 +peak memory allocated: 6168 MiB reserved: 6190 MiB +ema:applying EMA weights +DIAGNOSTIC post_ema val_loss:2.4290 val_bpb:1.4386 eval_time:48197ms +Serialized model: 106289590 bytes +Code size: 101850 bytes +gptq:building non-banked model for Hessian collection... +gptq:generating autoregressive calibration data (64 seqs x 2048 tokens, temp=0.8)... +gptq:generated 64 sequences in 197.4s +gptq:collecting hessians from autoregressive data... +gptq:collected hessians for 68 layers (AR self-gen) +selective_prune: 10619991 ±1 candidates, unpruned=7.62MB target=15.9MB +selective_prune: already fits, no pruning needed +Serialized model int6+lzma: 7891708 bytes +Total submission size int6+lzma: 7993558 bytes +final_int6_roundtrip val_loss:2.6833 val_bpb:1.5892 eval_time:50353ms +final_int6_roundtrip_exact val_loss:2.68325752 val_bpb:1.58917624 +final_int6_sliding_window val_loss:2.6484 val_bpb:1.5685 stride:64 eval_time:1614131ms +final_int6_sliding_window_exact val_loss:2.64841024 val_bpb:1.56854187 +final_int8_zlib_roundtrip_exact val_loss:2.64841024 val_bpb:1.56854187 +[W401 17:20:05.212673150 AllocatorConfig.cpp:28] Warning: PYTORCH_CUDA_ALLOC_CONF is deprecated, use PYTORCH_ALLOC_CONF instead (function operator()) +[W401 17:20:05.626244199 AllocatorConfig.cpp:28] Warning: PYTORCH_CUDA_ALLOC_CONF is deprecated, use PYTORCH_ALLOC_CONF instead (function operator()) diff --git a/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/train_a2_calib320_seed314.log b/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/train_a2_calib320_seed314.log new file mode 100644 index 0000000000..eb66a43cf4 --- /dev/null +++ b/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/train_a2_calib320_seed314.log @@ -0,0 +1,63 @@ +logs/a2_calib320_seed314.txt +val_bpb:enabled tokenizer_kind=sentencepiece tokenizer_path=./data/tokenizers/fineweb_1024_bpe.model +train_loader:dataset:fineweb10B_sp1024 train_shards:1 +val_loader:shards pattern=./data/datasets/fineweb10B_sp1024/fineweb_val_*.bin tokens:62021632 +model_params:27067484 +mtp_num_heads:0 mtp_loss_weight:0.2 mtp_params:0 +XSA:last_11 active_layers:[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10] +world_size:1 grad_accum_steps:8 +sdp_backends:cudnn=False flash=True mem_efficient=False math=False +attention_mode:gqa num_heads:8 num_kv_heads:4 +tie_embeddings:True embed_lr:0.035 head_lr:0.0 matrix_lr:0.025 scalar_lr:0.025 +train_batch_tokens:196608 train_seq_len:2048 iterations:20000 warmup_steps:20 max_wallclock_seconds:600.000 +seed:314 +warmup_step:1/20 +warmup_step:2/20 +warmup_step:3/20 +warmup_step:4/20 +warmup_step:5/20 +warmup_step:6/20 +warmup_step:7/20 +warmup_step:8/20 +warmup_step:9/20 +warmup_step:10/20 +warmup_step:11/20 +warmup_step:12/20 +warmup_step:13/20 +warmup_step:14/20 +warmup_step:15/20 +warmup_step:16/20 +warmup_step:17/20 +warmup_step:18/20 +warmup_step:19/20 +warmup_step:20/20 +step:1/20000 train_loss:6.9300 train_time:522ms step_avg:522.32ms +step:2/20000 train_loss:8.7825 train_time:993ms step_avg:496.61ms +step:3/20000 train_loss:8.3310 train_time:1511ms step_avg:503.75ms +step:4/20000 train_loss:7.6516 train_time:2057ms step_avg:514.16ms +step:5/20000 train_loss:7.0478 train_time:2574ms step_avg:514.83ms +step:6/20000 train_loss:6.5838 train_time:3094ms step_avg:515.65ms +step:7/20000 train_loss:6.3306 train_time:3614ms step_avg:516.32ms +step:8/20000 train_loss:6.0445 train_time:4165ms step_avg:520.59ms +step:9/20000 train_loss:5.9842 train_time:4683ms step_avg:520.36ms +step:10/20000 train_loss:5.9277 train_time:5202ms step_avg:520.24ms +swa:start step:400 +step:500/20000 train_loss:2.5963 train_time:255534ms step_avg:511.07ms +late_qat:enabled step:571 scale:0.1498 +step:1000/20000 train_loss:2.4018 train_time:512153ms step_avg:512.15ms +step:1170/20000 val_loss:2.3623 val_bpb:1.3991 train_time:600298ms step_avg:513.08ms +stopping_early: wallclock_cap train_time:600298ms step:1170/20000 +peak memory allocated: 6168 MiB reserved: 6190 MiB +ema:applying EMA weights +DIAGNOSTIC post_ema val_loss:2.4279 val_bpb:1.4379 eval_time:48188ms +Serialized model: 106289590 bytes +Code size: 101850 bytes +gptq:building non-banked model for Hessian collection... +gptq:generating autoregressive calibration data (64 seqs x 2048 tokens, temp=0.8)... +gptq:generated 64 sequences in 197.2s +gptq:collecting hessians from autoregressive data... +gptq:collected hessians for 68 layers (AR self-gen) +selective_prune: 10652298 ±1 candidates, unpruned=7.65MB target=15.9MB +selective_prune: already fits, no pruning needed +Serialized model int6+lzma: 7915212 bytes +Total submission size int6+lzma: 8017062 bytes diff --git a/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/train_b1_block256_seed314.log b/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/train_b1_block256_seed314.log new file mode 100644 index 0000000000..71086afcf2 --- /dev/null +++ b/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/train_b1_block256_seed314.log @@ -0,0 +1,65 @@ +logs/b1_block256_seed314.txt +val_bpb:enabled tokenizer_kind=sentencepiece tokenizer_path=./data/tokenizers/fineweb_1024_bpe.model +train_loader:dataset:fineweb10B_sp1024 train_shards:1 +val_loader:shards pattern=./data/datasets/fineweb10B_sp1024/fineweb_val_*.bin tokens:62021632 +model_params:27067484 +mtp_num_heads:0 mtp_loss_weight:0.2 mtp_params:0 +XSA:last_11 active_layers:[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10] +world_size:1 grad_accum_steps:8 +sdp_backends:cudnn=False flash=True mem_efficient=False math=False +attention_mode:gqa num_heads:8 num_kv_heads:4 +tie_embeddings:True embed_lr:0.035 head_lr:0.0 matrix_lr:0.025 scalar_lr:0.025 +train_batch_tokens:196608 train_seq_len:2048 iterations:20000 warmup_steps:20 max_wallclock_seconds:600.000 +seed:314 +warmup_step:1/20 +warmup_step:2/20 +warmup_step:3/20 +warmup_step:4/20 +warmup_step:5/20 +warmup_step:6/20 +warmup_step:7/20 +warmup_step:8/20 +warmup_step:9/20 +warmup_step:10/20 +warmup_step:11/20 +warmup_step:12/20 +warmup_step:13/20 +warmup_step:14/20 +warmup_step:15/20 +warmup_step:16/20 +warmup_step:17/20 +warmup_step:18/20 +warmup_step:19/20 +warmup_step:20/20 +step:1/20000 train_loss:6.9300 train_time:520ms step_avg:520.00ms +step:2/20000 train_loss:8.7825 train_time:997ms step_avg:498.37ms +step:3/20000 train_loss:8.3348 train_time:1513ms step_avg:504.38ms +step:4/20000 train_loss:7.6556 train_time:2027ms step_avg:506.77ms +step:5/20000 train_loss:7.0426 train_time:2541ms step_avg:508.26ms +step:6/20000 train_loss:6.5764 train_time:3060ms step_avg:509.93ms +step:7/20000 train_loss:6.3234 train_time:3595ms step_avg:513.54ms +step:8/20000 train_loss:6.0391 train_time:4119ms step_avg:514.85ms +step:9/20000 train_loss:5.9826 train_time:4635ms step_avg:515.03ms +step:10/20000 train_loss:5.9280 train_time:5155ms step_avg:515.55ms +swa:start step:400 +step:500/20000 train_loss:2.5967 train_time:255243ms step_avg:510.49ms +late_qat:enabled step:572 scale:0.1499 +step:1000/20000 train_loss:2.3966 train_time:511740ms step_avg:511.74ms +step:1171/20000 val_loss:2.3620 val_bpb:1.3989 train_time:600323ms step_avg:512.66ms +stopping_early: wallclock_cap train_time:600323ms step:1171/20000 +peak memory allocated: 6168 MiB reserved: 6190 MiB +ema:applying EMA weights +DIAGNOSTIC post_ema val_loss:2.4274 val_bpb:1.4376 eval_time:48183ms +Serialized model: 106289590 bytes +Code size: 101850 bytes +gptq:building non-banked model for Hessian collection... +gptq:generating autoregressive calibration data (64 seqs x 2048 tokens, temp=0.8)... +gptq:generated 64 sequences in 197.2s +gptq:collecting hessians from autoregressive data... +gptq:collected hessians for 68 layers (AR self-gen) +selective_prune: 10655605 ±1 candidates, unpruned=7.65MB target=15.9MB +selective_prune: already fits, no pruning needed +Serialized model int6+lzma: 7916960 bytes +Total submission size int6+lzma: 8018810 bytes +final_int6_roundtrip val_loss:2.6813 val_bpb:1.5880 eval_time:50370ms +final_int6_roundtrip_exact val_loss:2.68133118 val_bpb:1.58803536 diff --git a/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/train_ctrl_seed314.log b/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/train_ctrl_seed314.log new file mode 100644 index 0000000000..930d3a8219 --- /dev/null +++ b/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/train_ctrl_seed314.log @@ -0,0 +1,96 @@ +logs/ctrl_seed314.txt +val_bpb:enabled tokenizer_kind=sentencepiece tokenizer_path=./data/tokenizers/fineweb_1024_bpe.model +train_loader:dataset:fineweb10B_sp1024 train_shards:1 +val_loader:shards pattern=./data/datasets/fineweb10B_sp1024/fineweb_val_*.bin tokens:62021632 +model_params:27067484 +mtp_num_heads:0 mtp_loss_weight:0.2 mtp_params:0 +XSA:last_11 active_layers:[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10] +world_size:1 grad_accum_steps:8 +sdp_backends:cudnn=False flash=True mem_efficient=False math=False +attention_mode:gqa num_heads:8 num_kv_heads:4 +tie_embeddings:True embed_lr:0.035 head_lr:0.0 matrix_lr:0.025 scalar_lr:0.025 +train_batch_tokens:196608 train_seq_len:2048 iterations:20000 warmup_steps:20 max_wallclock_seconds:600.000 +seed:314 +warmup_step:1/20 +warmup_step:2/20 +warmup_step:3/20 +warmup_step:4/20 +warmup_step:5/20 +warmup_step:6/20 +warmup_step:7/20 +warmup_step:8/20 +warmup_step:9/20 +warmup_step:10/20 +warmup_step:11/20 +warmup_step:12/20 +warmup_step:13/20 +warmup_step:14/20 +warmup_step:15/20 +warmup_step:16/20 +warmup_step:17/20 +warmup_step:18/20 +warmup_step:19/20 +warmup_step:20/20 +step:1/20000 train_loss:6.9300 train_time:550ms step_avg:550.16ms +step:2/20000 train_loss:8.7825 train_time:1023ms step_avg:511.26ms +step:3/20000 train_loss:8.3534 train_time:1538ms step_avg:512.69ms +step:4/20000 train_loss:7.6774 train_time:2057ms step_avg:514.27ms +step:5/20000 train_loss:7.0621 train_time:2576ms step_avg:515.11ms +step:6/20000 train_loss:6.5916 train_time:3093ms step_avg:515.51ms +step:7/20000 train_loss:6.3355 train_time:3776ms step_avg:539.43ms +step:8/20000 train_loss:6.0481 train_time:4293ms step_avg:536.60ms +step:9/20000 train_loss:5.9874 train_time:4822ms step_avg:535.75ms +step:10/20000 train_loss:5.9259 train_time:5339ms step_avg:533.93ms +swa:start step:400 +step:500/20000 train_loss:2.5956 train_time:256246ms step_avg:512.49ms +late_qat:enabled step:567 scale:0.1500 +step:1000/20000 train_loss:2.4014 train_time:513333ms step_avg:513.33ms +step:1168/20000 val_loss:2.3636 val_bpb:1.3998 train_time:600275ms step_avg:513.93ms +stopping_early: wallclock_cap train_time:600275ms step:1168/20000 +peak memory allocated: 6168 MiB reserved: 6190 MiB +ema:applying EMA weights +DIAGNOSTIC post_ema val_loss:2.4292 val_bpb:1.4387 eval_time:48166ms +Serialized model: 106289590 bytes +Code size: 101850 bytes +gptq:building non-banked model for Hessian collection... +gptq:generating autoregressive calibration data (64 seqs x 2048 tokens, temp=0.8)... +gptq:generated 64 sequences in 197.4s +gptq:collecting hessians from autoregressive data... +gptq:collected hessians for 68 layers (AR self-gen) +selective_prune: 10633344 ±1 candidates, unpruned=7.63MB target=15.9MB +selective_prune: already fits, no pruning needed +Serialized model int6+lzma: 7900108 bytes +Total submission size int6+lzma: 8001958 bytes +final_int6_roundtrip val_loss:2.6829 val_bpb:1.5890 eval_time:61200ms +final_int6_roundtrip_exact val_loss:2.68292629 val_bpb:1.58898007 +W0401 16:35:31.680000 5135 torch/distributed/elastic/agent/server/api.py:725] Received 15 death signal, shutting down workers +W0401 16:35:31.681000 5135 torch/distributed/elastic/multiprocessing/api.py:908] Sending process 5172 closing signal SIGTERM +Traceback (most recent call last): + File "/usr/local/bin/torchrun", line 7, in + sys.exit(main()) + ^^^^^^ + File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/multiprocessing/errors/__init__.py", line 357, in wrapper + return f(*args, **kwargs) + ^^^^^^^^^^^^^^^^^^ + File "/usr/local/lib/python3.12/dist-packages/torch/distributed/run.py", line 936, in main + run(args) + File "/usr/local/lib/python3.12/dist-packages/torch/distributed/run.py", line 927, in run + elastic_launch( + File "/usr/local/lib/python3.12/dist-packages/torch/distributed/launcher/api.py", line 156, in __call__ + return launch_agent(self._config, self._entrypoint, list(args)) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ + File "/usr/local/lib/python3.12/dist-packages/torch/distributed/launcher/api.py", line 284, in launch_agent + result = agent.run() + ^^^^^^^^^^^ + File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/metrics/api.py", line 138, in wrapper + result = f(*args, **kwargs) + ^^^^^^^^^^^^^^^^^^ + File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/agent/server/api.py", line 717, in run + result = self._invoke_run(role) + ^^^^^^^^^^^^^^^^^^^^^^ + File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/agent/server/api.py", line 881, in _invoke_run + time.sleep(monitor_interval) + File "/usr/local/lib/python3.12/dist-packages/torch/distributed/elastic/multiprocessing/api.py", line 85, in _terminate_process_handler + raise SignalException(f"Process {os.getpid()} got signal: {sigval}", sigval=sigval) +torch.distributed.elastic.multiprocessing.api.SignalException: Process 5135 got signal: 15 +[W401 16:35:32.505266161 AllocatorConfig.cpp:28] Warning: PYTORCH_CUDA_ALLOC_CONF is deprecated, use PYTORCH_ALLOC_CONF instead (function operator()) diff --git a/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/train_gpt.py b/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/train_gpt.py new file mode 100644 index 0000000000..72c213f638 --- /dev/null +++ b/records/track_non_record_16mb/2026-04-01_Wolfie8935_4090_ValCalib_ablations/train_gpt.py @@ -0,0 +1,2135 @@ +from __future__ import annotations +import copy +import glob +import io +import lzma +import math +import os +import random +import subprocess +import sys +import time +import uuid +import zlib +from pathlib import Path +try: + import zstandard + _COMPRESSOR = "zstd" +except ImportError: + _COMPRESSOR = "zlib" +import numpy as np +import sentencepiece as spm +import torch +import torch.distributed as dist +import torch.nn.functional as F +from torch import Tensor, nn +from torch.nn.parallel import DistributedDataParallel as DDP +from flash_attn_interface import flash_attn_func as flash_attn_3_func +class Hyperparameters: + data_path = os.environ.get("DATA_PATH", "./data/datasets/fineweb10B_sp1024") + train_files = os.path.join(data_path, "fineweb_train_*.bin") + val_files = os.path.join(data_path, "fineweb_val_*.bin") + tokenizer_path = os.environ.get("TOKENIZER_PATH", "./data/tokenizers/fineweb_1024_bpe.model") + run_id = os.environ.get("RUN_ID", str(uuid.uuid4())) + seed = int(os.environ.get("SEED", 1337)) + val_batch_size = int(os.environ.get("VAL_BATCH_SIZE", 524_288)) + val_loss_every = int(os.environ.get("VAL_LOSS_EVERY", 4000)) + train_log_every = int(os.environ.get("TRAIN_LOG_EVERY", 500)) + iterations = int(os.environ.get("ITERATIONS", 20000)) + warmdown_iters = int(os.environ.get("WARMDOWN_ITERS", 3500)) + warmup_steps = int(os.environ.get("WARMUP_STEPS", 20)) + train_batch_tokens = int(os.environ.get("TRAIN_BATCH_TOKENS", 786_432)) + train_seq_len = int(os.environ.get("TRAIN_SEQ_LEN", 2048)) + eval_seq_len = int(os.environ.get("EVAL_SEQ_LEN", 2048)) + max_wallclock_seconds = float(os.environ.get("MAX_WALLCLOCK_SECONDS", 600.0)) + qk_gain_init = float(os.environ.get("QK_GAIN_INIT", 1.5)) + vocab_size = int(os.environ.get("VOCAB_SIZE", 1024)) + num_layers = int(os.environ.get("NUM_LAYERS", 11)) + num_kv_heads = int(os.environ.get("NUM_KV_HEADS", 4)) + model_dim = int(os.environ.get("MODEL_DIM", 512)) + num_heads = int(os.environ.get("NUM_HEADS", 8)) + mlp_mult = float(os.environ.get("MLP_MULT", 3.0)) + tie_embeddings = bool(int(os.environ.get("TIE_EMBEDDINGS", "1"))) + rope_base = float(os.environ.get("ROPE_BASE", 10000.0)) + logit_softcap = float(os.environ.get("LOGIT_SOFTCAP", 30.0)) + embed_lr = float(os.environ.get("EMBED_LR", 0.6)) + head_lr = float(os.environ.get("HEAD_LR", 0.008)) + tied_embed_lr = float(os.environ.get("TIED_EMBED_LR", 0.035)) + tied_embed_init_std = float(os.environ.get("TIED_EMBED_INIT_STD", 0.005)) + matrix_lr = float(os.environ.get("MATRIX_LR", 0.025)) + scalar_lr = float(os.environ.get("SCALAR_LR", 0.025)) + muon_momentum = float(os.environ.get("MUON_MOMENTUM", 0.99)) + muon_backend_steps = int(os.environ.get("MUON_BACKEND_STEPS", 5)) + muon_momentum_warmup_start = float(os.environ.get("MUON_MOMENTUM_WARMUP_START", 0.92)) + muon_momentum_warmup_steps = int(os.environ.get("MUON_MOMENTUM_WARMUP_STEPS", 1500)) + beta1 = float(os.environ.get("BETA1", 0.9)) + beta2 = float(os.environ.get("BETA2", 0.95)) + adam_eps = float(os.environ.get("ADAM_EPS", 1e-8)) + grad_clip_norm = float(os.environ.get("GRAD_CLIP_NORM", 0.3)) + eval_stride = int(os.environ.get("EVAL_STRIDE", 64)) + mtp_num_heads = int(os.environ.get("MTP_NUM_HEADS", 0)) + mtp_loss_weight = float(os.environ.get("MTP_LOSS_WEIGHT", 0.2)) + muon_beta2 = float(os.environ.get("MUON_BETA2", 0.95)) + swa_enabled = bool(int(os.environ.get("SWA_ENABLED", "1"))) + swa_every = int(os.environ.get("SWA_EVERY", 50)) + lawa_enabled = bool(int(os.environ.get("LAWA_ENABLED", "0"))) + lawa_k = int(os.environ.get("LAWA_K", 10)) + lawa_freq = int(os.environ.get("LAWA_FREQ", 100)) + muon_wd = float(os.environ.get("MUON_WD", 0.04)) + adam_wd = float(os.environ.get("ADAM_WD", 0.04)) + qat_enabled = bool(int(os.environ.get("QAT_ENABLED", "0"))) + bigram_vocab_size = int(os.environ.get("BIGRAM_VOCAB_SIZE", 2048)) + bigram_dim = int(os.environ.get("BIGRAM_DIM", 128)) + trigram_enabled = bool(int(os.environ.get("TRIGRAM", "0"))) # TrigramHash (off by default, risky) + xsa_last_n = int(os.environ.get("XSA_LAST_N", 11)) # XSA on ALL layers (our novel contribution) + rope_dims = int(os.environ.get("ROPE_DIMS", 16)) + ln_scale = bool(int(os.environ.get("LN_SCALE", "1"))) + dtg_enabled = bool(int(os.environ.get("DTG_ENABLED", "0"))) + late_qat_threshold = float(os.environ.get("LATE_QAT_THRESHOLD", 0.15)) + ve_enabled = bool(int(os.environ.get("VE_ENABLED", "1"))) + ve_dim = int(os.environ.get("VE_DIM", 128)) + ve_layers = os.environ.get("VE_LAYERS", "9,10") + gated_attention = bool(int(os.environ.get("GATED_ATTENTION", "0"))) + value_residual = bool(int(os.environ.get("VALUE_RESIDUAL", "0"))) # VRL with sigmoid gates (off by default, risky) + # GPTQ calibration + gptq_calib_batches = int(os.environ.get("GPTQ_CALIB_BATCHES", 256)) + gptq_block_size = int(os.environ.get("GPTQ_BLOCK_SIZE", 128)) + +# --- Batched Newton-Schulz orthogonalization --- + +def zeropower_via_newtonschulz5(G: Tensor, steps: int = 5, eps: float = 1e-7) -> Tensor: + """Batched Newton-Schulz orthogonalization. G: (B,M,N) or (M,N).""" + a, b, c = (3.4445, -4.7750, 2.0315) + was_2d = G.ndim == 2 + if was_2d: + G = G.unsqueeze(0) + X = G.bfloat16() + transposed = X.size(-2) > X.size(-1) + if transposed: + X = X.mT + X = X / (X.norm(dim=(-2, -1), keepdim=True) + eps) + for _ in range(steps): + A = X @ X.mT + B = b * A + c * (A @ A) + X = a * X + B @ X + if transposed: + X = X.mT + if was_2d: + X = X.squeeze(0) + return X + +# --- Parallel Muon optimizer --- + +class Muon(torch.optim.Optimizer): + """Parallel Muon: post-backward reduce-scatter -> local NS5 -> all-gather. + + No DDP for bank params. After backward, this optimizer: + 1. Launches async reduce-scatter for all banks (biggest first) + 2. Returns control so Adam can step on small params while RS is in-flight + 3. Waits for each RS, runs local NS5 on the shard, launches async all-gather + 4. Each all-gather overlaps with next bank's NS5 + """ + def __init__(self, params, lr: float, momentum: float, backend_steps: int, + nesterov: bool = True, weight_decay: float = 0.0): + super().__init__( + params, + dict(lr=lr, momentum=momentum, backend_steps=backend_steps, + nesterov=nesterov, weight_decay=weight_decay), + ) + self._built = False + + def _build(self): + self._distributed = dist.is_available() and dist.is_initialized() + self._world_size = dist.get_world_size() if self._distributed else 1 + self._rank = dist.get_rank() if self._distributed else 0 + ws = self._world_size + + self._bank_meta = [] + for group in self.param_groups: + for p in group["params"]: + B = p.shape[0] + padded_B = ((B + ws - 1) // ws) * ws + shard_B = padded_B // ws + tail = p.shape[1:] + dev = p.device + self._bank_meta.append({ + 'p': p, + 'B': B, + 'padded_grad': torch.zeros(padded_B, *tail, device=dev, dtype=torch.bfloat16), + 'shard': torch.zeros(shard_B, *tail, device=dev, dtype=torch.bfloat16), + 'shard_mom': torch.zeros(shard_B, *tail, device=dev, dtype=torch.bfloat16), + 'full_update': torch.zeros(padded_B, *tail, device=dev, dtype=torch.bfloat16), + 'scale': max(1, p.shape[-2] / p.shape[-1]) ** 0.5, + }) + # Sort by size descending -- launch biggest reduce-scatters first + self._bank_meta.sort(key=lambda m: -m['p'].numel()) + self._built = True + + def launch_reduce_scatters(self): + """Phase 1: launch async reduce-scatter for all banks. Call right after backward.""" + if not self._built: + self._build() + if not self._distributed: + return + self._rs_futures = [] + for m in self._bank_meta: + p = m['p'] + if p.grad is None: + self._rs_futures.append(None) + continue + pg = m['padded_grad'] + pg[:m['B']].copy_(p.grad.bfloat16()) + if pg.shape[0] > m['B']: + pg[m['B']:].zero_() + fut = dist.reduce_scatter_tensor(m['shard'], pg, op=dist.ReduceOp.AVG, async_op=True) + self._rs_futures.append(fut) + + @torch.no_grad() + def step(self, closure=None): + """Phase 3: wait for RS, local NS5, all-gather. Call AFTER Adam steps.""" + loss = None + if closure is not None: + with torch.enable_grad(): + loss = closure() + + if not self._built: + self._build() + + for group in self.param_groups: + lr = group["lr"] + momentum = group["momentum"] + backend_steps = group["backend_steps"] + nesterov = group["nesterov"] + wd = group.get("weight_decay", 0.0) + + prev_ag_handle = None + prev_m = None + + sharded = self._distributed and hasattr(self, '_rs_futures') + + for i, m in enumerate(self._bank_meta): + p = m['p'] + if p.grad is None: + continue + + if prev_ag_handle is not None: + prev_ag_handle.wait() + pp = prev_m['p'] + upd = prev_m['full_update'][:prev_m['B']] + if wd > 0.0: + pp.data.mul_(1.0 - lr * wd) + pp.add_(upd.to(dtype=pp.dtype), alpha=-lr * prev_m['scale']) + + if sharded and self._rs_futures[i] is not None: + self._rs_futures[i].wait() + g = m['shard'] + buf = m['shard_mom'] + else: + g = p.grad.bfloat16() + state = self.state[p] + if "momentum_buffer" not in state: + state["momentum_buffer"] = torch.zeros_like(g) + buf = state["momentum_buffer"] + + buf.mul_(momentum).add_(g) + if nesterov: + update = g.add(buf, alpha=momentum) + else: + update = buf + + update = zeropower_via_newtonschulz5(update, steps=backend_steps) + + if sharded: + prev_ag_handle = dist.all_gather_into_tensor( + m['full_update'], update, async_op=True) + prev_m = m + else: + if wd > 0.0: + p.data.mul_(1.0 - lr * wd) + p.add_(update.to(dtype=p.dtype), alpha=-lr * m['scale']) + + if prev_ag_handle is not None: + prev_ag_handle.wait() + pp = prev_m['p'] + upd = prev_m['full_update'][:prev_m['B']] + if wd > 0.0: + pp.data.mul_(1.0 - lr * wd) + pp.add_(upd.to(dtype=pp.dtype), alpha=-lr * prev_m['scale']) + + if hasattr(self, '_rs_futures'): + del self._rs_futures + + return loss + +# --- Tokenizer evaluation helpers --- + +def build_sentencepiece_luts( + sp: spm.SentencePieceProcessor, vocab_size: int, device: torch.device +) -> tuple[Tensor, Tensor, Tensor]: + sp_vocab_size = int(sp.vocab_size()) + table_size = max(sp_vocab_size, vocab_size) + base_bytes_np = np.zeros((table_size,), dtype=np.int16) + has_leading_space_np = np.zeros((table_size,), dtype=np.bool_) + is_boundary_token_np = np.ones((table_size,), dtype=np.bool_) + for token_id in range(sp_vocab_size): + if sp.is_control(token_id) or sp.is_unknown(token_id) or sp.is_unused(token_id): + continue + is_boundary_token_np[token_id] = False + if sp.is_byte(token_id): + base_bytes_np[token_id] = 1 + continue + piece = sp.id_to_piece(token_id) + if piece.startswith("\u2581"): + has_leading_space_np[token_id] = True + piece = piece[1:] + base_bytes_np[token_id] = len(piece.encode("utf-8")) + return ( + torch.tensor(base_bytes_np, dtype=torch.int16, device=device), + torch.tensor(has_leading_space_np, dtype=torch.bool, device=device), + torch.tensor(is_boundary_token_np, dtype=torch.bool, device=device), + ) +def load_validation_tokens(pattern: str, seq_len: int) -> Tensor: + files = [Path(p) for p in sorted(glob.glob(pattern))] + if not files: + raise FileNotFoundError(f"No files found for pattern: {pattern}") + tokens = torch.cat([load_data_shard(file) for file in files]).contiguous() + usable = ((tokens.numel() - 1) // seq_len) * seq_len + if usable <= 0: + raise ValueError(f"Validation split is too short for TRAIN_SEQ_LEN={seq_len}") + return tokens[: usable + 1] +def eval_val( + args: Hyperparameters, + model: nn.Module, + rank: int, + world_size: int, + device: torch.device, + grad_accum_steps: int, + val_tokens: Tensor, + base_bytes_lut: Tensor, + has_leading_space_lut: Tensor, + is_boundary_token_lut: Tensor, + eval_seq_len: int | None = None, +) -> tuple[float, float]: + seq_len = eval_seq_len or args.train_seq_len + local_batch_tokens = args.val_batch_size // (world_size * grad_accum_steps) + if local_batch_tokens < seq_len: + raise ValueError( + "VAL_BATCH_SIZE must provide at least one sequence per rank; " + f"got VAL_BATCH_SIZE={args.val_batch_size}, WORLD_SIZE={world_size}, " + f"GRAD_ACCUM_STEPS={grad_accum_steps}, seq_len={seq_len}" + ) + local_batch_seqs = local_batch_tokens // seq_len + total_seqs = (val_tokens.numel() - 1) // seq_len + seq_start = (total_seqs * rank) // world_size + seq_end = (total_seqs * (rank + 1)) // world_size + val_loss_sum = torch.zeros((), device=device, dtype=torch.float64) + val_token_count = torch.zeros((), device=device, dtype=torch.float64) + val_byte_count = torch.zeros((), device=device, dtype=torch.float64) + model.eval() + with torch.inference_mode(): + for batch_seq_start in range(seq_start, seq_end, local_batch_seqs): + batch_seq_end = min(batch_seq_start + local_batch_seqs, seq_end) + raw_start = batch_seq_start * seq_len + raw_end = batch_seq_end * seq_len + 1 + local = val_tokens[raw_start:raw_end].to(device=device, dtype=torch.int64, non_blocking=True) + x = local[:-1].reshape(-1, seq_len) + y = local[1:].reshape(-1, seq_len) + with torch.autocast(device_type="cuda", dtype=torch.bfloat16, enabled=True): + batch_loss = model(x, y).detach() + batch_token_count = float(y.numel()) + val_loss_sum += batch_loss.to(torch.float64) * batch_token_count + val_token_count += batch_token_count + prev_ids = x.reshape(-1) + tgt_ids = y.reshape(-1) + token_bytes = base_bytes_lut[tgt_ids].to(dtype=torch.int16) + token_bytes += (has_leading_space_lut[tgt_ids] & ~is_boundary_token_lut[prev_ids]).to(dtype=torch.int16) + val_byte_count += token_bytes.to(torch.float64).sum() + if dist.is_available() and dist.is_initialized(): + dist.all_reduce(val_loss_sum, op=dist.ReduceOp.SUM) + dist.all_reduce(val_token_count, op=dist.ReduceOp.SUM) + dist.all_reduce(val_byte_count, op=dist.ReduceOp.SUM) + val_loss = val_loss_sum / val_token_count + bits_per_token = val_loss.item() / math.log(2.0) + tokens_per_byte = val_token_count.item() / val_byte_count.item() + model.train() + return float(val_loss.item()), float(bits_per_token * tokens_per_byte) + +# --- Quantization helpers --- + +CONTROL_TENSOR_NAME_PATTERNS = tuple( + pattern + for pattern in os.environ.get( + "CONTROL_TENSOR_NAME_PATTERNS", + "attn_scale,attn_scales,mlp_scale,mlp_scales,resid_mix,resid_mixes,q_gain,skip_weight,skip_weights,smear,dtg_gate,ve_layer_scales,ve_shared.scale,attn_gate,vr_lambda", + ).split(",") + if pattern +) +INT8_KEEP_FLOAT_FP32_NAME_PATTERNS = tuple( + pattern + for pattern in os.environ.get( + "INT8_KEEP_FLOAT_FP32_NAME_PATTERNS", + ",".join(CONTROL_TENSOR_NAME_PATTERNS), + ).split(",") + if pattern +) +INT8_KEEP_FLOAT_MAX_NUMEL = 65_536 +INT8_KEEP_FLOAT_STORE_DTYPE = torch.float16 +INT8_PER_ROW_SCALE_DTYPE = torch.float16 +INT8_CLIP_PERCENTILE = 99.99984 +INT8_CLIP_Q = INT8_CLIP_PERCENTILE / 100.0 +def tensor_nbytes(t: Tensor) -> int: + return int(t.numel()) * int(t.element_size()) +def keep_float_tensor(name: str, t: Tensor, passthrough_orig_dtypes: dict[str, str]) -> Tensor: + if any(pattern in name for pattern in INT8_KEEP_FLOAT_FP32_NAME_PATTERNS): + return t.float().contiguous() + if t.dtype in {torch.float32, torch.bfloat16}: + passthrough_orig_dtypes[name] = str(t.dtype).removeprefix("torch.") + return t.to(dtype=INT8_KEEP_FLOAT_STORE_DTYPE).contiguous() + return t +def quantize_float_tensor(t: Tensor) -> tuple[Tensor, Tensor]: + t32 = t.float() + if t32.ndim == 2: + clip_abs = ( + torch.quantile(t32.abs(), INT8_CLIP_Q, dim=1) + if t32.numel() + else torch.empty((t32.shape[0],), dtype=torch.float32) + ) + clipped = torch.maximum(torch.minimum(t32, clip_abs[:, None]), -clip_abs[:, None]) + scale = (clip_abs / 127.0).clamp_min(1.0 / 127.0) + q = torch.clamp(torch.round(clipped / scale[:, None]), -127, 127).to(torch.int8).contiguous() + return q, scale.to(dtype=INT8_PER_ROW_SCALE_DTYPE).contiguous() + clip_abs = float(torch.quantile(t32.abs().flatten(), INT8_CLIP_Q).item()) if t32.numel() else 0.0 + scale = torch.tensor(clip_abs / 127.0 if clip_abs > 0 else 1.0, dtype=torch.float32) + q = torch.clamp(torch.round(torch.clamp(t32, -clip_abs, clip_abs) / scale), -127, 127).to(torch.int8).contiguous() + return q, scale +def quantize_state_dict_int8(state_dict: dict[str, Tensor]): + quantized: dict[str, Tensor] = {} + scales: dict[str, Tensor] = {} + dtypes: dict[str, str] = {} + passthrough: dict[str, Tensor] = {} + passthrough_orig_dtypes: dict[str, str] = {} + qmeta: dict[str, dict[str, object]] = {} + stats = dict.fromkeys( + ("param_count", "num_tensors", "num_float_tensors", "num_nonfloat_tensors", "baseline_tensor_bytes", "int8_payload_bytes"), + 0, + ) + for name, tensor in state_dict.items(): + t = tensor.detach().to("cpu").contiguous() + stats["param_count"] += int(t.numel()) + stats["num_tensors"] += 1 + stats["baseline_tensor_bytes"] += tensor_nbytes(t) + if not t.is_floating_point(): + stats["num_nonfloat_tensors"] += 1 + passthrough[name] = t + stats["int8_payload_bytes"] += tensor_nbytes(t) + continue + if t.numel() <= INT8_KEEP_FLOAT_MAX_NUMEL: + kept = keep_float_tensor(name, t, passthrough_orig_dtypes) + passthrough[name] = kept + stats["int8_payload_bytes"] += tensor_nbytes(kept) + continue + stats["num_float_tensors"] += 1 + q, s = quantize_float_tensor(t) + if s.ndim > 0: + qmeta[name] = {"scheme": "per_row", "axis": 0} + quantized[name] = q + scales[name] = s + dtypes[name] = str(t.dtype).removeprefix("torch.") + stats["int8_payload_bytes"] += tensor_nbytes(q) + tensor_nbytes(s) + obj: dict[str, object] = { + "__quant_format__": "int8_clean_per_row_v1", + "quantized": quantized, + "scales": scales, + "dtypes": dtypes, + "passthrough": passthrough, + } + if qmeta: + obj["qmeta"] = qmeta + if passthrough_orig_dtypes: + obj["passthrough_orig_dtypes"] = passthrough_orig_dtypes + return obj, stats +def dequantize_state_dict_int8(obj: dict[str, object]) -> dict[str, Tensor]: + out: dict[str, Tensor] = {} + qmeta = obj.get("qmeta", {}) + passthrough_orig_dtypes = obj.get("passthrough_orig_dtypes", {}) + for name, q in obj["quantized"].items(): + dtype = getattr(torch, obj["dtypes"][name]) + s = obj["scales"][name] + if qmeta.get(name, {}).get("scheme") == "per_row" or s.ndim > 0: + s = s.to(dtype=torch.float32) + out[name] = (q.float() * s.view(q.shape[0], *([1] * (q.ndim - 1)))).to(dtype=dtype).contiguous() + else: + scale = float(s.item()) + out[name] = (q.float() * scale).to(dtype=dtype).contiguous() + for name, t in obj["passthrough"].items(): + out_t = t.detach().to("cpu").contiguous() + orig_dtype = passthrough_orig_dtypes.get(name) + if isinstance(orig_dtype, str): + out_t = out_t.to(dtype=getattr(torch, orig_dtype)).contiguous() + out[name] = out_t + return out + +# --- Data loading --- + +def load_data_shard(file: Path) -> Tensor: + header_bytes = 256 * np.dtype(" None: + self.file_idx = (self.file_idx + 1) % len(self.files) + self.tokens = load_data_shard(self.files[self.file_idx]) + self.pos = 0 + def take(self, n: int) -> Tensor: + chunks: list[Tensor] = [] + remaining = n + while remaining > 0: + avail = self.tokens.numel() - self.pos + if avail <= 0: + self._advance_file() + continue + k = min(remaining, avail) + chunks.append(self.tokens[self.pos : self.pos + k]) + self.pos += k + remaining -= k + return chunks[0] if len(chunks) == 1 else torch.cat(chunks) +class DistributedTokenLoader: + def __init__(self, pattern: str, rank: int, world_size: int, device: torch.device): + self.rank = rank + self.world_size = world_size + self.device = device + self.stream = TokenStream(pattern) + def next_batch(self, global_tokens: int, seq_len: int, grad_accum_steps: int) -> tuple[Tensor, Tensor]: + local_tokens = global_tokens // (self.world_size * grad_accum_steps) + per_rank_span = local_tokens + 1 + chunk = self.stream.take(per_rank_span * self.world_size) + start = self.rank * per_rank_span + local = chunk[start : start + per_rank_span].to(dtype=torch.int64) + x = local[:-1].reshape(-1, seq_len) + y = local[1:].reshape(-1, seq_len) + return x.to(self.device, non_blocking=True), y.to(self.device, non_blocking=True) + +# --- Transformer modules --- + +class RMSNorm(nn.Module): + def __init__(self, eps: float | None = None): + super().__init__() + self.eps = eps + def forward(self, x: Tensor) -> Tensor: + return F.rms_norm(x, (x.size(-1),), eps=self.eps) +class CastedLinear(nn.Linear): + _qat_enabled: bool = False + def forward(self, x: Tensor) -> Tensor: + w = self.weight.to(x.dtype) + if CastedLinear._qat_enabled and self.training and w.ndim == 2: + with torch.no_grad(): + w32 = self.weight.float() + row_max = w32.abs().amax(dim=1) + scale = (row_max / 31.0).clamp_min(1.0 / 31.0) + w_q = (torch.clamp(torch.round(w32 / scale[:, None]), -32, 31) * scale[:, None]).to(x.dtype) + w = w + (w_q - w).detach() + bias = self.bias.to(x.dtype) if self.bias is not None else None + return F.linear(x, w, bias) +def restore_low_dim_params_to_fp32(module: nn.Module) -> None: + with torch.no_grad(): + for name, param in module.named_parameters(): + if (param.ndim < 2 or any(pattern in name for pattern in CONTROL_TENSOR_NAME_PATTERNS)) and param.dtype != torch.float32: + param.data = param.data.float() +class Rotary(nn.Module): + def __init__(self, dim: int, base: float = 10000.0, train_seq_len: int = 1024, rope_dims: int = 0): + super().__init__() + self.dim = dim + self.base = base + self.train_seq_len = train_seq_len + self.rope_dims = rope_dims if rope_dims > 0 else dim + inv_freq = 1.0 / (base ** (torch.arange(0, self.rope_dims, 2, dtype=torch.float32) / self.rope_dims)) + self.register_buffer("inv_freq", inv_freq, persistent=False) + self._seq_len_cached = 0 + self._cos_cached: Tensor | None = None + self._sin_cached: Tensor | None = None + def forward(self, seq_len: int, device: torch.device, dtype: torch.dtype) -> tuple[Tensor, Tensor]: + if ( + self._cos_cached is None + or self._sin_cached is None + or self._seq_len_cached != seq_len + or self._cos_cached.device != device + ): + rd = self.rope_dims + if seq_len > self.train_seq_len: + scale = seq_len / self.train_seq_len + new_base = self.base * (scale ** (rd / (rd - 2))) + inv_freq = 1.0 / (new_base ** (torch.arange(0, rd, 2, dtype=torch.float32, device=device) / rd)) + else: + inv_freq = self.inv_freq.to(device) + t = torch.arange(seq_len, device=device, dtype=inv_freq.dtype) + freqs = torch.outer(t, inv_freq) + self._cos_cached = freqs.cos()[None, :, None, :] + self._sin_cached = freqs.sin()[None, :, None, :] + self._seq_len_cached = seq_len + return self._cos_cached.to(dtype=dtype), self._sin_cached.to(dtype=dtype) +def apply_rotary_emb(x: Tensor, cos: Tensor, sin: Tensor, rope_dims: int = 0) -> Tensor: + if rope_dims > 0 and rope_dims < x.size(-1): + x_rope, x_pass = x[..., :rope_dims], x[..., rope_dims:] + half = rope_dims // 2 + x1, x2 = x_rope[..., :half], x_rope[..., half:] + x_rope = torch.cat((x1 * cos + x2 * sin, x1 * (-sin) + x2 * cos), dim=-1) + return torch.cat((x_rope, x_pass), dim=-1) + half = x.size(-1) // 2 + x1, x2 = x[..., :half], x[..., half:] + return torch.cat((x1 * cos + x2 * sin, x1 * (-sin) + x2 * cos), dim=-1) + +class CausalSelfAttention(nn.Module): + def __init__( + self, + dim: int, + num_heads: int, + num_kv_heads: int, + rope_base: float, + qk_gain_init: float, + gated_attention: bool = False, + value_residual: bool = False, + ): + super().__init__() + if dim % num_heads != 0: + raise ValueError("model_dim must be divisible by num_heads") + if num_heads % num_kv_heads != 0: + raise ValueError("num_heads must be divisible by num_kv_heads") + self.num_heads = num_heads + self.num_kv_heads = num_kv_heads + self.head_dim = dim // num_heads + if self.head_dim % 2 != 0: + raise ValueError("head_dim must be even for RoPE") + # No CastedLinear -- weights come from banks + self.q_gain = nn.Parameter(torch.full((num_heads,), qk_gain_init, dtype=torch.float32)) + self.rope_dims = 0 # set by GPT.__init__ for partial RoPE + self.rotary = Rotary(self.head_dim, base=rope_base, train_seq_len=1024) + self.use_xsa = False # set by GPT.__init__ for deep layers only + # Gated attention and value residual (non-banked small params) + self.gated_attention = gated_attention + if gated_attention: + self.attn_gate = nn.Linear(dim, num_heads, bias=True) + nn.init.zeros_(self.attn_gate.weight) + nn.init.constant_(self.attn_gate.bias, 4.0) + self.value_residual = value_residual + if value_residual: + self.vrl_alpha = nn.Parameter(torch.zeros(1, dtype=torch.float32)) # sigmoid gate (PR #569 style) + def _xsa_efficient(self, y: Tensor, v: Tensor) -> Tensor: + """Efficient XSA: subtract self-value projection via GQA-aware reshape (no repeat_interleave). + y: [B, T, H, D], v: [B, T, Hkv, D]. H must be divisible by Hkv.""" + B, T, H, D = y.shape + Hkv = v.size(-2) + group = H // Hkv + y_g = y.reshape(B, T, Hkv, group, D) # [B, T, Hkv, group, D] + vn = F.normalize(v, dim=-1).unsqueeze(-2) # [B, T, Hkv, 1, D] -- broadcast ready + proj = (y_g * vn).sum(dim=-1, keepdim=True) * vn + return (y_g - proj).reshape(B, T, H, D) + def forward(self, x: Tensor, q_w: Tensor, k_w: Tensor, v_w: Tensor, out_w: Tensor, v_embed: Tensor | None = None, v0: Tensor | None = None) -> tuple[Tensor, Tensor | None]: + bsz, seqlen, dim = x.shape + q = F.linear(x, q_w.to(x.dtype)).reshape(bsz, seqlen, self.num_heads, self.head_dim) + k = F.linear(x, k_w.to(x.dtype)).reshape(bsz, seqlen, self.num_kv_heads, self.head_dim) + v = F.linear(x, v_w.to(x.dtype)) + if v_embed is not None: + v = v + v_embed + v = v.reshape(bsz, seqlen, self.num_kv_heads, self.head_dim) + raw_v = v if self.value_residual else None + if self.value_residual and v0 is not None: + alpha = torch.sigmoid(self.vrl_alpha.to(dtype=v.dtype)) + v = v + alpha * v0 # sigmoid-gated residual (PR #569 style) + q = F.rms_norm(q, (q.size(-1),)) + k = F.rms_norm(k, (k.size(-1),)) + cos, sin = self.rotary(seqlen, x.device, q.dtype) + q = apply_rotary_emb(q, cos, sin, self.rope_dims) + k = apply_rotary_emb(k, cos, sin, self.rope_dims) + q = q * self.q_gain.to(dtype=q.dtype)[None, None, :, None] + y = flash_attn_3_func(q, k, v, causal=True) + if self.use_xsa: + y = self._xsa_efficient(y, v) + if self.gated_attention: + # gate shape: (bsz, seqlen, num_heads) -> (bsz, seqlen, num_heads, 1) for B,T,H,D layout + gate = torch.sigmoid(self.attn_gate(x)).unsqueeze(-1) + y = y * gate + y = y.reshape(bsz, seqlen, dim) + return F.linear(y, out_w.to(x.dtype)), raw_v + +class SmearGate(nn.Module): + def __init__(self, dim: int): + super().__init__() + self.gate = nn.Parameter(torch.zeros(dim, dtype=torch.float32)) + def forward(self, x: Tensor) -> Tensor: + g = torch.sigmoid(self.gate.to(dtype=x.dtype))[None, None, :] + x_prev = torch.cat([torch.zeros_like(x[:, :1]), x[:, :-1]], dim=1) + return (1 - g) * x + g * x_prev + +class BigramHashEmbedding(nn.Module): + def __init__(self, bigram_vocab_size: int, bigram_dim: int, model_dim: int, trigram: bool = False): + super().__init__() + self.bigram_vocab_size = bigram_vocab_size + self._trigram = trigram + self.embed = nn.Embedding(bigram_vocab_size, bigram_dim) + nn.init.zeros_(self.embed.weight) + self.proj = CastedLinear(bigram_dim, model_dim, bias=False) if bigram_dim != model_dim else None + if self.proj is not None: + nn.init.zeros_(self.proj.weight) + self.scale = nn.Parameter(torch.tensor(0.05, dtype=torch.float32)) + def bigram_hash(self, tokens: Tensor) -> Tensor: + t = tokens.to(torch.int32) + mod = self.bigram_vocab_size - 1 + out = torch.empty_like(t) + out[..., 0] = mod + out[..., 1:] = torch.bitwise_xor(36313 * t[..., 1:], 27191 * t[..., :-1]) % mod + return out.long() + def trigram_hash(self, tokens: Tensor) -> Tensor: + """Hash (t-2, t-1, t) trigrams into same embedding table. Zero extra params.""" + t = tokens.to(torch.int32) + mod = self.bigram_vocab_size - 1 + out = torch.empty_like(t) + out[..., :2] = mod + out[..., 2:] = (36313 * t[..., 2:] ^ 27191 * t[..., 1:-1] ^ 51497 * t[..., :-2]) % mod + return out.long() + def forward(self, token_ids: Tensor) -> Tensor: + h = self.embed(self.bigram_hash(token_ids)) + if self._trigram: + h = h + self.embed(self.trigram_hash(token_ids)) + if self.proj is not None: + h = self.proj(h) + return h * self.scale.to(dtype=h.dtype) + +class ValueEmbedding(nn.Module): + """Reinject token identity into attention values at specific layers. + Each table maps vocab tokens to a low-dim embedding, projected to model_dim.""" + def __init__(self, vocab_size: int, ve_dim: int, model_dim: int): + super().__init__() + self.embed = nn.Embedding(vocab_size, ve_dim) + nn.init.normal_(self.embed.weight, std=0.01) + self.proj = CastedLinear(ve_dim, model_dim, bias=False) if ve_dim != model_dim else None + if self.proj is not None: + nn.init.zeros_(self.proj.weight) + self.scale = nn.Parameter(torch.tensor(0.1, dtype=torch.float32)) + def forward(self, token_ids: Tensor) -> Tensor: + h = self.embed(token_ids) + if self.proj is not None: + h = self.proj(h) + return h * self.scale.to(dtype=h.dtype) + +class MLP(nn.Module): + def __init__(self, dim: int, mlp_mult: int): + super().__init__() + # No CastedLinear -- weights come from banks + def forward(self, x: Tensor, up_w: Tensor, down_w: Tensor) -> Tensor: + x = F.leaky_relu(F.linear(x, up_w.to(x.dtype)), negative_slope=0.5) + return F.linear(x.square(), down_w.to(x.dtype)) + +class Block(nn.Module): + def __init__( + self, + dim: int, + num_heads: int, + num_kv_heads: int, + mlp_mult: int, + rope_base: float, + qk_gain_init: float, + layer_idx: int = 0, + ln_scale: bool = False, + dtg: bool = False, + gated_attention: bool = False, + value_residual: bool = False, + ): + super().__init__() + self.attn_norm = RMSNorm() + self.mlp_norm = RMSNorm() + self.attn = CausalSelfAttention(dim, num_heads, num_kv_heads, rope_base, qk_gain_init, + gated_attention=gated_attention, value_residual=value_residual) + self.mlp = MLP(dim, mlp_mult) + self.attn_scale = nn.Parameter(torch.ones(dim, dtype=torch.float32)) + self.mlp_scale = nn.Parameter(torch.ones(dim, dtype=torch.float32)) + self.resid_mix = nn.Parameter(torch.stack((torch.ones(dim), torch.zeros(dim))).float()) + self.ln_scale_factor = 1.0 / math.sqrt(layer_idx + 1) if ln_scale else 1.0 + if dtg: + self.dtg_gate = nn.Linear(dim, 1, bias=True) + nn.init.zeros_(self.dtg_gate.weight) + nn.init.constant_(self.dtg_gate.bias, 2.0) + else: + self.dtg_gate = None + def forward(self, x: Tensor, x0: Tensor, q_w: Tensor, k_w: Tensor, v_w: Tensor, out_w: Tensor, up_w: Tensor, down_w: Tensor, v_embed: Tensor | None = None, v0: Tensor | None = None) -> tuple[Tensor, Tensor | None]: + mix = self.resid_mix.to(dtype=x.dtype) + x_in = mix[0][None, None, :] * x + mix[1][None, None, :] * x0 + attn_out, raw_v = self.attn(self.attn_norm(x_in) * self.ln_scale_factor, q_w, k_w, v_w, out_w, v_embed=v_embed, v0=v0) + x_out = x_in + self.attn_scale.to(dtype=x_in.dtype)[None, None, :] * attn_out + x_out = x_out + self.mlp_scale.to(dtype=x_out.dtype)[None, None, :] * self.mlp(self.mlp_norm(x_out) * self.ln_scale_factor, up_w, down_w) + if self.dtg_gate is not None: + gate = torch.sigmoid(self.dtg_gate(x_in.detach())) + x_out = x_in + gate * (x_out - x_in) + return x_out, raw_v + +class GPT(nn.Module): + def __init__( + self, + vocab_size: int, + num_layers: int, + model_dim: int, + num_heads: int, + num_kv_heads: int, + mlp_mult: int, + tie_embeddings: bool, + tied_embed_init_std: float, + logit_softcap: float, + rope_base: float, + qk_gain_init: float, + mtp_num_heads: int = 0, + mtp_loss_weight: float = 0.1, + bigram_vocab_size: int = 0, + bigram_dim: int = 128, + xsa_last_n: int = 0, + rope_dims: int = 0, + ln_scale: bool = False, + dtg: bool = False, + ve_enabled: bool = False, + ve_dim: int = 128, + ve_layers: str = "9,10", + gated_attention: bool = False, + value_residual: bool = False, + ): + super().__init__() + self._ve_target_dim = num_kv_heads * (model_dim // num_heads) # kv_dim for value projection + if logit_softcap <= 0.0: + raise ValueError(f"logit_softcap must be positive, got {logit_softcap}") + self.tie_embeddings = tie_embeddings + self.tied_embed_init_std = tied_embed_init_std + self.logit_softcap = logit_softcap + self.value_residual = value_residual + self.mtp_num_heads = mtp_num_heads + self.mtp_loss_weight = mtp_loss_weight + self.tok_emb = nn.Embedding(vocab_size, model_dim) + self.bigram = BigramHashEmbedding(bigram_vocab_size, bigram_dim, model_dim, trigram=bool(int(os.environ.get("TRIGRAM", "0")))) if bigram_vocab_size > 0 else None + self.smear = SmearGate(model_dim) + self.num_encoder_layers = num_layers // 2 + self.num_decoder_layers = num_layers - self.num_encoder_layers + self.num_skip_weights = min(self.num_encoder_layers, self.num_decoder_layers) + self.skip_weights = nn.Parameter(torch.ones(self.num_skip_weights, model_dim, dtype=torch.float32)) + # Parameter banks: contiguous 3D tensors for batched optimizer + head_dim = model_dim // num_heads + kv_dim = num_kv_heads * head_dim + mlp_dim = int(mlp_mult * model_dim) + self.num_layers = num_layers + self.qo_bank = nn.Parameter(torch.empty(2 * num_layers, model_dim, model_dim)) + self.kv_bank = nn.Parameter(torch.empty(2 * num_layers, kv_dim, model_dim)) + self.mlp_up_bank = nn.Parameter(torch.empty(num_layers, mlp_dim, model_dim)) + self.mlp_down_bank = nn.Parameter(torch.empty(num_layers, model_dim, mlp_dim)) + self.blocks = nn.ModuleList( + [ + Block( + model_dim, + num_heads, + num_kv_heads, + mlp_mult, + rope_base, + qk_gain_init, + layer_idx=i, + ln_scale=ln_scale, + dtg=dtg, + gated_attention=gated_attention, + value_residual=value_residual, + ) + for i in range(num_layers) + ] + ) + if rope_dims > 0: + head_dim = model_dim // num_heads + for block in self.blocks: + block.attn.rope_dims = rope_dims + block.attn.rotary = Rotary(head_dim, base=rope_base, train_seq_len=1024, rope_dims=rope_dims) + self.ve_layer_indices = [int(x) for x in ve_layers.split(",") if x.strip()] if ve_enabled else [] + kv_dim_ve = self._ve_target_dim + if self.ve_layer_indices: + self.ve_shared = ValueEmbedding(vocab_size, ve_dim, kv_dim_ve) + self.ve_layer_scales = nn.ParameterList( + [nn.Parameter(torch.ones(1, dtype=torch.float32)) for _ in self.ve_layer_indices] + ) + else: + self.ve_shared = None + self.ve_layer_scales = nn.ParameterList() + self.value_embeds = nn.ModuleList() # keep empty for compat + self.final_norm = RMSNorm() + self.lm_head = None if tie_embeddings else CastedLinear(model_dim, vocab_size, bias=False) + if self.lm_head is not None: + self.lm_head._zero_init = True + self.mtp_heads = nn.ModuleList( + [CastedLinear(model_dim, vocab_size, bias=False) for _ in range(mtp_num_heads)] + ) + for head in self.mtp_heads: + head._zero_init = True + if xsa_last_n > 0: + for i in range(max(0, num_layers - xsa_last_n), num_layers): + self.blocks[i].attn.use_xsa = True + self._init_weights() + def _init_weights(self) -> None: + if self.tie_embeddings: + nn.init.normal_(self.tok_emb.weight, mean=0.0, std=self.tied_embed_init_std) + n = self.num_layers + proj_scale = 1.0 / math.sqrt(2 * n) + # Init banks: orthogonal, with proj layers scaled down and out/down zero-init + for i in range(n): + nn.init.orthogonal_(self.qo_bank.data[i], gain=1.0) # Q + nn.init.zeros_(self.qo_bank.data[n + i]) # Out (zero init) + nn.init.orthogonal_(self.kv_bank.data[i], gain=1.0) # K + nn.init.orthogonal_(self.kv_bank.data[n + i], gain=1.0) # V + nn.init.orthogonal_(self.mlp_up_bank.data[i], gain=1.0) # MLP up + nn.init.zeros_(self.mlp_down_bank.data[i]) # MLP down (zero init) + # Scale proj layers (out_proj and mlp_down are "proj" layers) + self.qo_bank.data[n + i].mul_(proj_scale) + self.mlp_down_bank.data[i].mul_(proj_scale) + # Init remaining nn.Linear modules (bigram proj, mtp heads, lm_head) + for name, module in self.named_modules(): + if isinstance(module, nn.Linear): + if getattr(module, "_zero_init", False): + nn.init.zeros_(module.weight) + elif module.weight.ndim == 2 and module.weight.shape[0] >= 64 and module.weight.shape[1] >= 64: + nn.init.orthogonal_(module.weight, gain=1.0) + def _get_ve(self, layer_idx: int, input_ids: Tensor, ve_cache: dict | None = None) -> Tensor | None: + """Get value embedding for a specific layer using shared table + per-layer scale.""" + if self.ve_shared is None or layer_idx not in self.ve_layer_indices: + return None + if ve_cache is not None and 've' not in ve_cache: + ve_cache['ve'] = self.ve_shared(input_ids) + ve_base = ve_cache['ve'] if ve_cache is not None else self.ve_shared(input_ids) + ve_idx = self.ve_layer_indices.index(layer_idx) + return ve_base * self.ve_layer_scales[ve_idx].to(dtype=ve_base.dtype) + def forward(self, input_ids: Tensor, target_ids: Tensor) -> Tensor: + n = self.num_layers + x = self.tok_emb(input_ids) + if self.bigram is not None: + x = x + self.bigram(input_ids) + x = F.rms_norm(x, (x.size(-1),)) + x = self.smear(x) + x0 = x + v0 = None + skips: list[Tensor] = [] + ve_cache: dict = {} + for i in range(self.num_encoder_layers): + ve = self._get_ve(i, input_ids, ve_cache) + x, raw_v = self.blocks[i](x, x0, + self.qo_bank[i], self.kv_bank[i], self.kv_bank[n + i], + self.qo_bank[n + i], self.mlp_up_bank[i], self.mlp_down_bank[i], + v_embed=ve, v0=v0) + if v0 is None and raw_v is not None: + v0 = raw_v + skips.append(x) + for i in range(self.num_decoder_layers): + bi = self.num_encoder_layers + i + if skips: + x = x + self.skip_weights[i].to(dtype=x.dtype)[None, None, :] * skips.pop() + ve = self._get_ve(bi, input_ids, ve_cache) + x, _ = self.blocks[bi](x, x0, + self.qo_bank[bi], self.kv_bank[bi], self.kv_bank[n + bi], + self.qo_bank[n + bi], self.mlp_up_bank[bi], self.mlp_down_bank[bi], + v_embed=ve, v0=v0) + x = self.final_norm(x) + x_flat = x.reshape(-1, x.size(-1)) + targets = target_ids.reshape(-1) + if self.tie_embeddings: + logits_proj = F.linear(x_flat, self.tok_emb.weight) + else: + if self.lm_head is None: + raise RuntimeError("lm_head is required when tie_embeddings=False") + logits_proj = self.lm_head(x_flat) + logits = self.logit_softcap * torch.tanh(logits_proj / self.logit_softcap) + main_loss = F.cross_entropy(logits.float(), targets, reduction="mean") + if self.training and self.mtp_num_heads > 0 and self.mtp_loss_weight > 0.0: + _, seqlen, dim = x.shape + mtp_loss_sum = x.new_zeros(()) + mtp_loss_count = 0 + for k, mtp_head in enumerate(self.mtp_heads): + valid_t = seqlen - (k + 1) + if valid_t <= 0: + continue + mtp_hidden = x[:, :valid_t, :].reshape(-1, dim) + mtp_targets = target_ids[:, k + 1 :].reshape(-1) + mtp_logits_proj = mtp_head(mtp_hidden) + mtp_logits = self.logit_softcap * torch.tanh(mtp_logits_proj / self.logit_softcap) + mtp_loss_sum = mtp_loss_sum + F.cross_entropy(mtp_logits.float(), mtp_targets, reduction="mean") + mtp_loss_count += 1 + if mtp_loss_count > 0: + main_loss = main_loss + self.mtp_loss_weight * (mtp_loss_sum / mtp_loss_count) + return main_loss + def forward_logits(self, input_ids: Tensor) -> Tensor: + """Return logits (bsz, seq_len, vocab) without computing loss.""" + n = self.num_layers + x = self.tok_emb(input_ids) + if self.bigram is not None: + x = x + self.bigram(input_ids) + x = F.rms_norm(x, (x.size(-1),)) + x = self.smear(x) + x0 = x + v0 = None + skips: list[Tensor] = [] + ve_cache: dict = {} + for i in range(self.num_encoder_layers): + ve = self._get_ve(i, input_ids, ve_cache) + x, raw_v = self.blocks[i](x, x0, + self.qo_bank[i], self.kv_bank[i], self.kv_bank[n + i], + self.qo_bank[n + i], self.mlp_up_bank[i], self.mlp_down_bank[i], + v_embed=ve, v0=v0) + if v0 is None and raw_v is not None: + v0 = raw_v + skips.append(x) + for i in range(self.num_decoder_layers): + bi = self.num_encoder_layers + i + if skips: + x = x + self.skip_weights[i].to(dtype=x.dtype)[None, None, :] * skips.pop() + ve = self._get_ve(bi, input_ids, ve_cache) + x, _ = self.blocks[bi](x, x0, + self.qo_bank[bi], self.kv_bank[bi], self.kv_bank[n + bi], + self.qo_bank[n + bi], self.mlp_up_bank[bi], self.mlp_down_bank[bi], + v_embed=ve, v0=v0) + x = self.final_norm(x) + if self.tie_embeddings: + logits_proj = F.linear(x, self.tok_emb.weight) + else: + logits_proj = self.lm_head(x) + return self.logit_softcap * torch.tanh(logits_proj / self.logit_softcap) + +# --- Sliding window evaluation --- + +def eval_val_sliding( + args: Hyperparameters, + base_model: nn.Module, + rank: int, + world_size: int, + device: torch.device, + val_tokens: Tensor, + base_bytes_lut: Tensor, + has_leading_space_lut: Tensor, + is_boundary_token_lut: Tensor, + stride: int, + batch_seqs: int = 32, + eval_seq_len: int | None = None, +) -> tuple[float, float]: + """Sliding window evaluation: each token scored with maximum context.""" + seq_len = eval_seq_len or args.train_seq_len + total_tokens = val_tokens.numel() - 1 + window_starts = [ws for ws in range(0, total_tokens, stride) + if min(ws + seq_len, total_tokens) - ws >= 1] + total_windows = len(window_starts) + my_s = (total_windows * rank) // world_size + my_e = (total_windows * (rank + 1)) // world_size + my_windows = window_starts[my_s:my_e] + loss_sum = torch.zeros((), device=device, dtype=torch.float64) + token_count = torch.zeros((), device=device, dtype=torch.float64) + byte_count = torch.zeros((), device=device, dtype=torch.float64) + base_model.eval() + compiled_logits = torch.compile(base_model.forward_logits, dynamic=False, fullgraph=True) + with torch.inference_mode(): + for bi in range(0, len(my_windows), batch_seqs): + batch_ws = my_windows[bi:bi + batch_seqs] + bsz = len(batch_ws) + x_batch = torch.zeros(bsz, seq_len, dtype=torch.int64, device=device) + y_batch = torch.zeros(bsz, seq_len, dtype=torch.int64, device=device) + wlens: list[int] = [] + for i, ws in enumerate(batch_ws): + end = min(ws + seq_len, total_tokens) + wlen = end - ws + wlens.append(wlen) + chunk = val_tokens[ws:end + 1].to(dtype=torch.int64, device=device) + x_batch[i, :wlen] = chunk[:-1] + y_batch[i, :wlen] = chunk[1:] + with torch.autocast(device_type="cuda", dtype=torch.bfloat16): + logits = compiled_logits(x_batch) + nll = F.cross_entropy( + logits.reshape(-1, logits.size(-1)).float(), + y_batch.reshape(-1), + reduction="none", + ).reshape(bsz, seq_len) + for i, ws in enumerate(batch_ws): + wlen = wlens[i] + s = 0 if ws == 0 else max(wlen - stride, 0) + scored_nll = nll[i, s:wlen].to(torch.float64) + loss_sum += scored_nll.sum() + token_count += float(wlen - s) + tgt = y_batch[i, s:wlen] + prev = x_batch[i, s:wlen] + tb = base_bytes_lut[tgt].to(torch.float64) + tb += (has_leading_space_lut[tgt] & ~is_boundary_token_lut[prev]).to(torch.float64) + byte_count += tb.sum() + if dist.is_available() and dist.is_initialized(): + dist.all_reduce(loss_sum, op=dist.ReduceOp.SUM) + dist.all_reduce(token_count, op=dist.ReduceOp.SUM) + dist.all_reduce(byte_count, op=dist.ReduceOp.SUM) + val_loss = (loss_sum / token_count).item() + bits_per_token = val_loss / math.log(2.0) + tokens_per_byte = token_count.item() / byte_count.item() + base_model.train() + return val_loss, bits_per_token * tokens_per_byte + + +def generate_autoregressive_calib(model, device, num_seqs=64, seq_len=2048, + vocab_size=1024, temperature=0.8, batch_size=8, seed=42): + """Generate sequences autoregressively from the model for GPTQ calibration. + No external data accessed — fully self-contained.""" + model.eval() + rng = torch.Generator(device=device) + rng.manual_seed(seed) + all_tokens = [] + with torch.inference_mode(), torch.autocast(device_type="cuda", dtype=torch.bfloat16): + for batch_start in range(0, num_seqs, batch_size): + bs = min(batch_size, num_seqs - batch_start) + tokens = torch.randint(0, vocab_size, (bs, 1), device=device, generator=rng) + for pos in range(seq_len - 1): + logits = model.forward_logits(tokens) + next_logit = logits[:, -1, :] + probs = torch.softmax(next_logit / temperature, dim=-1) + next_tok = torch.multinomial(probs, 1, generator=rng) + tokens = torch.cat([tokens, next_tok], dim=1) + for i in range(bs): + all_tokens.append(tokens[i:i+1]) + return all_tokens + + +def collect_hessians_from_tokens(hessian_model, token_seqs, device): + """Collect H = X^T X from pre-generated token sequences.""" + hessians = {} + hooks = [] + for name, module in hessian_model.named_modules(): + if isinstance(module, CastedLinear): + param_name = name + ".weight" + cols = module.weight.shape[1] + hessians[param_name] = torch.zeros(cols, cols, dtype=torch.float32, device='cpu') + def make_hook(pname): + def hook_fn(module, input, output): + x = input[0].detach().float() + if x.ndim == 3: + x = x.reshape(-1, x.shape[-1]) + hessians[pname] += (x.T @ x).cpu() + return hook_fn + h = module.register_forward_hook(make_hook(param_name)) + hooks.append(h) + hessian_model.eval() + with torch.inference_mode(), torch.autocast(device_type="cuda", dtype=torch.bfloat16): + for seq in token_seqs: + x = seq[:, :-1].to(device) + y = seq[:, 1:].to(device) + hessian_model(x, y) + for h in hooks: + h.remove() + num_batches = len(token_seqs) + for name in hessians: + H = hessians[name] + H /= num_batches + damp = 0.01 * torch.diag(H).mean().clamp_min(1e-6) + H += damp * torch.eye(H.shape[0]) + hessians[name] = H + return hessians + + +# --- GPTQ-lite int6 quantization --- + +def _classify_param(name: str) -> str: + if "tok_emb" in name or "lm_head" in name: + return "embed" + if ".mlp." in name: + return "mlp" + if ".attn." in name or (".proj." in name and ".mlp." not in name): + return "attn" + return "other" +def quantize_int6_per_row(t: Tensor, clip_range: int = 31) -> tuple[Tensor, Tensor]: + t32 = t.float() + if t32.ndim == 2: + best_q, best_s, best_err = None, None, float('inf') + for pct in [0.9990, 0.9995, 0.9999, 0.99999, 1.0]: + if pct < 1.0: + row_clip = torch.quantile(t32.abs(), pct, dim=1) + else: + row_clip = t32.abs().amax(dim=1) + s = (row_clip / clip_range).clamp_min(1.0 / clip_range).to(torch.float16) + q = torch.clamp(torch.round(t32 / s.float()[:, None]), -clip_range, clip_range).to(torch.int8) + recon = q.float() * s.float()[:, None] + err = (t32 - recon).pow(2).mean().item() + if err < best_err: + best_q, best_s, best_err = q, s, err + return best_q, best_s + amax = t32.abs().max().item() + scale = torch.tensor(amax / clip_range if amax > 0 else 1.0, dtype=torch.float16) + q = torch.clamp(torch.round(t32 / scale.float()), -clip_range, clip_range).to(torch.int8) + return q, scale + +def quantize_int6_gptq(weight, hessian=None, clip_range=31, block_size=128): + """Full GPTQ: Hessian-aware int6 quantization with Cholesky error compensation. + If hessian is None, falls back to percentile search.""" + t32 = weight.float() + if t32.ndim != 2 or hessian is None: + return _quantize_int6_percentile(t32, clip_range) + rows, cols = t32.shape + H = hessian.float().clone() + dead = torch.diag(H) == 0 + H[dead, dead] = 1 + damp = 0.01 * torch.mean(torch.diag(H)) + H[torch.arange(cols), torch.arange(cols)] += damp + perm = torch.argsort(torch.diag(H), descending=True) + inv_perm = torch.argsort(perm) + W = t32[:, perm].clone() + W[:, dead[perm]] = 0 + H = H[perm][:, perm] + Hinv = torch.linalg.cholesky(H) + Hinv = torch.cholesky_inverse(Hinv) + Hinv = torch.linalg.cholesky(Hinv, upper=True) + best_q = None; best_scale = None; best_err = float('inf') + for pct in [0.9990, 0.9995, 0.9999, 0.99999, 1.0]: + if pct < 1.0: + row_clip = torch.quantile(t32.abs(), pct, dim=1) + else: + row_clip = t32.abs().amax(dim=1) + s = (row_clip / clip_range).clamp_min(1.0 / clip_range).to(torch.float16) + sf = s.float() + Q = torch.zeros_like(W, dtype=torch.int8) + W_work = W.clone() + for i1 in range(0, cols, block_size): + i2 = min(i1 + block_size, cols) + count = i2 - i1 + W1 = W_work[:, i1:i2].clone() + Q1 = torch.zeros(rows, count, dtype=torch.int8) + Err1 = torch.zeros(rows, count) + Hinv1 = Hinv[i1:i2, i1:i2] + for i in range(count): + w = W1[:, i] + d = Hinv1[i, i] + q = torch.clamp(torch.round(w / sf), -clip_range, clip_range).to(torch.int8) + Q1[:, i] = q + err = (w - q.float() * sf) / d + W1[:, i:] -= err.unsqueeze(1) * Hinv1[i, i:].unsqueeze(0) + Err1[:, i] = err + Q[:, i1:i2] = Q1 + if i2 < cols: + W_work[:, i2:] -= Err1 @ Hinv[i1:i2, i2:] + recon = Q.float() * sf[:, None] + mse = (W - recon).pow(2).mean().item() + if mse < best_err: + best_q, best_scale, best_err = Q, s, mse + best_q = best_q[:, inv_perm] + return best_q, best_scale + +def _quantize_int6_percentile(t32, clip_range=31): + """Fallback: percentile search (for 1D or no-Hessian cases).""" + if t32.ndim == 2: + best_q, best_s, best_err = None, None, float('inf') + for pct in [0.9990, 0.9995, 0.9999, 0.99999, 1.0]: + if pct < 1.0: + row_clip = torch.quantile(t32.abs(), pct, dim=1) + else: + row_clip = t32.abs().amax(dim=1) + s = (row_clip / clip_range).clamp_min(1.0 / clip_range).to(torch.float16) + q = torch.clamp(torch.round(t32 / s.float()[:, None]), -clip_range, clip_range).to(torch.int8) + recon = q.float() * s.float()[:, None] + err = (t32 - recon).pow(2).mean().item() + if err < best_err: + best_q, best_s, best_err = q, s, err + return best_q, best_s + amax = t32.abs().max().item() + scale = torch.tensor(amax / clip_range if amax > 0 else 1.0, dtype=torch.float16) + q = torch.clamp(torch.round(t32 / scale.float()), -clip_range, clip_range).to(torch.int8) + return q, scale + +def _unbank_state_dict(sd: dict[str, Tensor], num_layers: int) -> dict[str, Tensor]: + """Convert 3D bank tensors into individual 2D tensors with standard names.""" + out: dict[str, Tensor] = {} + n = num_layers + for name, tensor in sd.items(): + if name == "qo_bank": + for i in range(n): + out[f"blocks.{i}.attn.c_q.weight"] = tensor[i] + out[f"blocks.{i}.attn.proj.weight"] = tensor[n + i] + elif name == "kv_bank": + for i in range(n): + out[f"blocks.{i}.attn.c_k.weight"] = tensor[i] + out[f"blocks.{i}.attn.c_v.weight"] = tensor[n + i] + elif name == "mlp_up_bank": + for i in range(n): + out[f"blocks.{i}.mlp.fc.weight"] = tensor[i] + elif name == "mlp_down_bank": + for i in range(n): + out[f"blocks.{i}.mlp.proj.weight"] = tensor[i] + else: + out[name] = tensor + return out + +def _rebank_state_dict(sd: dict[str, Tensor], num_layers: int, template_sd: dict[str, Tensor]) -> dict[str, Tensor]: + """Convert individual 2D tensors back into 3D bank tensors.""" + out: dict[str, Tensor] = {} + n = num_layers + # Reconstruct banks from individual weight keys + qo_slices = [None] * (2 * n) + kv_slices = [None] * (2 * n) + up_slices = [None] * n + down_slices = [None] * n + consumed = set() + for i in range(n): + qk = f"blocks.{i}.attn.c_q.weight" + if qk in sd: + qo_slices[i] = sd[qk] + consumed.add(qk) + ok = f"blocks.{i}.attn.proj.weight" + if ok in sd: + qo_slices[n + i] = sd[ok] + consumed.add(ok) + kk = f"blocks.{i}.attn.c_k.weight" + if kk in sd: + kv_slices[i] = sd[kk] + consumed.add(kk) + vk = f"blocks.{i}.attn.c_v.weight" + if vk in sd: + kv_slices[n + i] = sd[vk] + consumed.add(vk) + fk = f"blocks.{i}.mlp.fc.weight" + if fk in sd: + up_slices[i] = sd[fk] + consumed.add(fk) + dk = f"blocks.{i}.mlp.proj.weight" + if dk in sd: + down_slices[i] = sd[dk] + consumed.add(dk) + out["qo_bank"] = torch.stack(qo_slices).to(dtype=template_sd["qo_bank"].dtype) + out["kv_bank"] = torch.stack(kv_slices).to(dtype=template_sd["kv_bank"].dtype) + out["mlp_up_bank"] = torch.stack(up_slices).to(dtype=template_sd["mlp_up_bank"].dtype) + out["mlp_down_bank"] = torch.stack(down_slices).to(dtype=template_sd["mlp_down_bank"].dtype) + for name, tensor in sd.items(): + if name not in consumed: + out[name] = tensor + return out + +# --- Non-banked model for Hessian collection --- +# This mirrors the unbanked state dict keys: blocks.{i}.attn.c_q/c_k/c_v/proj, blocks.{i}.mlp.fc/proj + +class _HessianAttn(nn.Module): + """Non-banked attention with CastedLinear layers for Hessian hooks.""" + def __init__(self, dim, num_heads, num_kv_heads, rope_base, qk_gain_init): + super().__init__() + self.num_heads, self.num_kv_heads = num_heads, num_kv_heads + self.head_dim = dim // num_heads + kv_dim = num_kv_heads * self.head_dim + self.c_q = CastedLinear(dim, dim, bias=False) + self.c_k = CastedLinear(dim, kv_dim, bias=False) + self.c_v = CastedLinear(dim, kv_dim, bias=False) + self.proj = CastedLinear(dim, dim, bias=False) + self.q_gain = nn.Parameter(torch.full((num_heads,), qk_gain_init, dtype=torch.float32)) + self.rope_dims = 0 + self.rotary = Rotary(self.head_dim, base=rope_base, train_seq_len=1024) + self.use_xsa = False + def _xsa_efficient(self, y, v): + B, T, H, D = y.shape; Hkv = v.size(-2); group = H // Hkv + y_g = y.reshape(B, T, Hkv, group, D) + vn = F.normalize(v, dim=-1).unsqueeze(-2) + proj = (y_g * vn).sum(dim=-1, keepdim=True) * vn + return (y_g - proj).reshape(B, T, H, D) + def forward(self, x, v_embed=None): + bsz, seqlen, dim = x.shape + q = self.c_q(x).reshape(bsz, seqlen, self.num_heads, self.head_dim) + k = self.c_k(x).reshape(bsz, seqlen, self.num_kv_heads, self.head_dim) + v = self.c_v(x) + if v_embed is not None: + v = v + v_embed + v = v.reshape(bsz, seqlen, self.num_kv_heads, self.head_dim) + q = F.rms_norm(q, (q.size(-1),)) + k = F.rms_norm(k, (k.size(-1),)) + cos, sin = self.rotary(seqlen, x.device, q.dtype) + q = apply_rotary_emb(q, cos, sin, self.rope_dims) + k = apply_rotary_emb(k, cos, sin, self.rope_dims) + q = q * self.q_gain.to(dtype=q.dtype)[None, None, :, None] + y = flash_attn_3_func(q, k, v, causal=True) + if self.use_xsa: + y = self._xsa_efficient(y, v) + return self.proj(y.reshape(bsz, seqlen, dim)) + +class _HessianMLP(nn.Module): + """Non-banked MLP with CastedLinear layers for Hessian hooks.""" + def __init__(self, dim, mlp_mult): + super().__init__() + self.fc = CastedLinear(dim, int(mlp_mult * dim), bias=False) + self.proj = CastedLinear(int(mlp_mult * dim), dim, bias=False) + def forward(self, x): + return self.proj(F.leaky_relu(self.fc(x), negative_slope=0.5).square()) + +class _HessianBlock(nn.Module): + def __init__(self, dim, num_heads, num_kv_heads, mlp_mult, rope_base, qk_gain_init, layer_idx=0, ln_scale=False): + super().__init__() + self.attn_norm = RMSNorm() + self.mlp_norm = RMSNorm() + self.attn = _HessianAttn(dim, num_heads, num_kv_heads, rope_base, qk_gain_init) + self.mlp = _HessianMLP(dim, mlp_mult) + self.attn_scale = nn.Parameter(torch.ones(dim, dtype=torch.float32)) + self.mlp_scale = nn.Parameter(torch.ones(dim, dtype=torch.float32)) + self.resid_mix = nn.Parameter(torch.stack((torch.ones(dim), torch.zeros(dim))).float()) + self.ln_scale_factor = 1.0 / math.sqrt(layer_idx + 1) if ln_scale else 1.0 + def forward(self, x, x0, v_embed=None): + mix = self.resid_mix.to(dtype=x.dtype) + x_in = mix[0][None, None, :] * x + mix[1][None, None, :] * x0 + attn_out = self.attn(self.attn_norm(x_in) * self.ln_scale_factor, v_embed=v_embed) + x_out = x_in + self.attn_scale.to(dtype=x_in.dtype)[None, None, :] * attn_out + x_out = x_out + self.mlp_scale.to(dtype=x_out.dtype)[None, None, :] * self.mlp(self.mlp_norm(x_out) * self.ln_scale_factor) + return x_out + +class _HessianGPT(nn.Module): + """Non-banked GPT model matching unbanked state dict keys for Hessian collection.""" + def __init__(self, vocab_size, num_layers, model_dim, num_heads, num_kv_heads, + mlp_mult, tie_embeddings, logit_softcap, rope_base, qk_gain_init, + bigram_vocab_size=0, bigram_dim=128, xsa_last_n=0, + rope_dims=0, ln_scale=False, + ve_enabled=False, ve_dim=128, ve_layers="9,10"): + super().__init__() + self.tie_embeddings = tie_embeddings + self.logit_softcap = logit_softcap + self.num_layers = num_layers + self.tok_emb = nn.Embedding(vocab_size, model_dim) + self.bigram = BigramHashEmbedding(bigram_vocab_size, bigram_dim, model_dim, trigram=bool(int(os.environ.get("TRIGRAM", "0")))) if bigram_vocab_size > 0 else None + self.smear = SmearGate(model_dim) + self.num_encoder_layers = num_layers // 2 + self.num_decoder_layers = num_layers - self.num_encoder_layers + self.num_skip_weights = min(self.num_encoder_layers, self.num_decoder_layers) + self.skip_weights = nn.Parameter(torch.ones(self.num_skip_weights, model_dim, dtype=torch.float32)) + self.blocks = nn.ModuleList([ + _HessianBlock(model_dim, num_heads, num_kv_heads, mlp_mult, rope_base, qk_gain_init, + layer_idx=i, ln_scale=ln_scale) + for i in range(num_layers) + ]) + if rope_dims > 0: + head_dim = model_dim // num_heads + for block in self.blocks: + block.attn.rope_dims = rope_dims + block.attn.rotary = Rotary(head_dim, base=rope_base, train_seq_len=1024, rope_dims=rope_dims) + if xsa_last_n > 0: + for i in range(max(0, num_layers - xsa_last_n), num_layers): + self.blocks[i].attn.use_xsa = True + kv_dim = num_kv_heads * (model_dim // num_heads) + self.ve_layer_indices = [int(x) for x in ve_layers.split(",") if x.strip()] if ve_enabled else [] + if self.ve_layer_indices: + self.ve_shared = ValueEmbedding(vocab_size, ve_dim, kv_dim) + self.ve_layer_scales = nn.ParameterList([nn.Parameter(torch.ones(1, dtype=torch.float32)) for _ in self.ve_layer_indices]) + else: + self.ve_shared = None + self.ve_layer_scales = nn.ParameterList() + self.final_norm = RMSNorm() + self.lm_head = None if tie_embeddings else CastedLinear(model_dim, vocab_size, bias=False) + def _get_ve(self, layer_idx, input_ids, ve_cache): + if self.ve_shared is None or layer_idx not in self.ve_layer_indices: + return None + if 've' not in ve_cache: + ve_cache['ve'] = self.ve_shared(input_ids) + ve_idx = self.ve_layer_indices.index(layer_idx) + return ve_cache['ve'] * self.ve_layer_scales[ve_idx].to(dtype=ve_cache['ve'].dtype) + def forward(self, input_ids, target_ids): + x = self.tok_emb(input_ids) + if self.bigram is not None: + x = x + self.bigram(input_ids) + x = F.rms_norm(x, (x.size(-1),)) + x = self.smear(x) + x0 = x + skips = [] + ve_cache = {} + for i in range(self.num_encoder_layers): + ve = self._get_ve(i, input_ids, ve_cache) + x = self.blocks[i](x, x0, v_embed=ve) + skips.append(x) + for i in range(self.num_decoder_layers): + bi = self.num_encoder_layers + i + if skips: + x = x + self.skip_weights[i].to(dtype=x.dtype)[None, None, :] * skips.pop() + ve = self._get_ve(bi, input_ids, ve_cache) + x = self.blocks[bi](x, x0, v_embed=ve) + x = self.final_norm(x) + x_flat = x.reshape(-1, x.size(-1)) + targets = target_ids.reshape(-1) + logits_proj = F.linear(x_flat, self.tok_emb.weight) if self.tie_embeddings else self.lm_head(x_flat) + logits = self.logit_softcap * torch.tanh(logits_proj / self.logit_softcap) + return F.cross_entropy(logits.float(), targets, reduction="mean") + +def collect_hessians(hessian_model, train_loader, args, device, grad_accum_steps, num_batches=256): + """Run calibration batches through a non-banked model, collecting H = X^T X for each CastedLinear.""" + hessians = {} + hooks = [] + for name, module in hessian_model.named_modules(): + if isinstance(module, CastedLinear): + param_name = name + ".weight" + cols = module.weight.shape[1] + hessians[param_name] = torch.zeros(cols, cols, dtype=torch.float32, device='cpu') + def make_hook(pname): + def hook_fn(module, input, output): + x = input[0].detach().float() + if x.ndim == 3: + x = x.reshape(-1, x.shape[-1]) + hessians[pname] += (x.T @ x).cpu() + return hook_fn + h = module.register_forward_hook(make_hook(param_name)) + hooks.append(h) + hessian_model.eval() + with torch.inference_mode(), torch.autocast(device_type="cuda", dtype=torch.bfloat16): + for _ in range(num_batches): + x, y = train_loader.next_batch(args.train_batch_tokens, args.train_seq_len, grad_accum_steps) + hessian_model(x, y) + for h in hooks: + h.remove() + for name in hessians: + H = hessians[name] + H /= num_batches + damp = 0.01 * torch.diag(H).mean().clamp_min(1e-6) + H += damp * torch.eye(H.shape[0]) + hessians[name] = H + hessian_model.train() + return hessians + +def mixed_quantize_int6(state_dict: dict[str, Tensor], int6_cats: set[str], hessians: dict[str, Tensor] | None = None): + num_layers_total = max( + (int(k.split(".")[1]) for k in state_dict if k.startswith("blocks.")), + default=0, + ) + 1 + late_k_layers = set(range(num_layers_total - 2, num_layers_total)) + result: dict[str, Tensor] = {} + meta: dict[str, object] = {} + for name, tensor in state_dict.items(): + t = tensor.detach().cpu().contiguous() + cat = _classify_param(name) + if not t.is_floating_point() or t.numel() <= 65536: + result[name] = t.to(torch.float16) if t.is_floating_point() else t + meta[name] = "passthrough" + continue + if any(p in name for p in CONTROL_TENSOR_NAME_PATTERNS): + result[name] = t.float() + meta[name] = "passthrough_ctrl" + continue + if cat in int6_cats and t.ndim >= 1: + cr = 31 # int6 for all weights + H = hessians.get(name) if hessians else None + if H is not None: + q, s = quantize_int6_gptq(t, hessian=H, clip_range=cr) + else: + q, s = quantize_int6_per_row(t, clip_range=cr) + result[name + ".q"] = q + result[name + ".scale"] = s + meta[name] = {"type": "int6"} + else: + q, s = quantize_float_tensor(t) + result[name + ".q"] = q + result[name + ".scale"] = s + meta[name] = {"type": "int8"} + return result, meta +def dequantize_mixed_int6(result: dict[str, Tensor], meta: dict[str, object], + template_sd: dict[str, Tensor]) -> dict[str, Tensor]: + out: dict[str, Tensor] = {} + for name, orig in template_sd.items(): + info = meta.get(name) + if info is None: + continue + orig_dtype = orig.dtype + if info in ("passthrough", "passthrough_ctrl", "passthrough_fp16"): + t = result[name] + if t.dtype == torch.float16 and orig_dtype in (torch.float32, torch.bfloat16): + t = t.to(orig_dtype) + out[name] = t + continue + q, s = result[name + ".q"], result[name + ".scale"] + if s.ndim > 0: + out[name] = (q.float() * s.float().view(q.shape[0], *([1] * (q.ndim - 1)))).to(orig_dtype) + else: + out[name] = (q.float() * float(s.item())).to(orig_dtype) + return out + +# --- Training --- + +def main() -> None: + code = Path(__file__).read_text(encoding="utf-8") + args = Hyperparameters() + # zeropower_via_newtonschulz5 runs eagerly with bmm -- do NOT compile + distributed = "RANK" in os.environ and "WORLD_SIZE" in os.environ + rank = int(os.environ.get("RANK", "0")) + world_size = int(os.environ.get("WORLD_SIZE", "1")) + local_rank = int(os.environ.get("LOCAL_RANK", "0")) + if world_size <= 0: + raise ValueError(f"WORLD_SIZE must be positive, got {world_size}") + if 8 % world_size != 0: + raise ValueError(f"WORLD_SIZE={world_size} must divide 8 so grad_accum_steps stays integral") + grad_accum_steps = 8 // world_size + grad_scale = 1.0 / grad_accum_steps + if not torch.cuda.is_available(): + raise RuntimeError("CUDA is required") + device = torch.device("cuda", local_rank) + torch.cuda.set_device(device) + if distributed: + dist.init_process_group(backend="nccl", device_id=device) + dist.barrier() + master_process = rank == 0 + torch.backends.cuda.matmul.allow_tf32 = True + torch.backends.cudnn.allow_tf32 = True + from torch.backends.cuda import enable_cudnn_sdp, enable_flash_sdp, enable_math_sdp, enable_mem_efficient_sdp + enable_cudnn_sdp(False) + enable_flash_sdp(True) + enable_mem_efficient_sdp(False) + enable_math_sdp(False) + logfile = None + if master_process: + os.makedirs("logs", exist_ok=True) + logfile = f"logs/{args.run_id}.txt" + print(logfile) + def log0(msg: str, console: bool = True) -> None: + if not master_process: + return + if console: + print(msg) + if logfile is not None: + with open(logfile, "a", encoding="utf-8") as f: + print(msg, file=f) + log0(code, console=False) + log0("=" * 100, console=False) + log0(f"Running Python {sys.version}", console=False) + log0(f"Running PyTorch {torch.__version__}", console=False) + log0( + subprocess.run(["nvidia-smi"], stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True, check=False).stdout, + console=False, + ) + log0("=" * 100, console=False) + random.seed(args.seed) + np.random.seed(args.seed) + torch.manual_seed(args.seed) + torch.cuda.manual_seed_all(args.seed) + if not args.tokenizer_path.endswith(".model"): + raise ValueError(f"Script only setup for SentencePiece .model file: {args.tokenizer_path}") + sp = spm.SentencePieceProcessor(model_file=args.tokenizer_path) + if int(sp.vocab_size()) != args.vocab_size: + raise ValueError( + f"VOCAB_SIZE={args.vocab_size} does not match tokenizer vocab_size={int(sp.vocab_size())}" + ) + dataset_dir = Path(args.data_path).resolve() + actual_train_files = len(list(dataset_dir.glob("fineweb_train_*.bin"))) + effective_eval_seq_len = args.eval_seq_len if args.eval_seq_len > 0 else args.train_seq_len + val_seq_len = max(args.train_seq_len, effective_eval_seq_len) + val_tokens = load_validation_tokens(args.val_files, val_seq_len) + base_bytes_lut, has_leading_space_lut, is_boundary_token_lut = build_sentencepiece_luts( + sp, args.vocab_size, device + ) + log0(f"val_bpb:enabled tokenizer_kind=sentencepiece tokenizer_path={args.tokenizer_path}") + log0(f"train_loader:dataset:{dataset_dir.name} train_shards:{actual_train_files}") + log0(f"val_loader:shards pattern={args.val_files} tokens:{val_tokens.numel() - 1}") + CastedLinear._qat_enabled = args.qat_enabled + base_model = GPT( + vocab_size=args.vocab_size, + num_layers=args.num_layers, + model_dim=args.model_dim, + num_heads=args.num_heads, + num_kv_heads=args.num_kv_heads, + mlp_mult=args.mlp_mult, + tie_embeddings=args.tie_embeddings, + tied_embed_init_std=args.tied_embed_init_std, + logit_softcap=args.logit_softcap, + rope_base=args.rope_base, + qk_gain_init=args.qk_gain_init, + mtp_num_heads=args.mtp_num_heads, + mtp_loss_weight=args.mtp_loss_weight, + bigram_vocab_size=args.bigram_vocab_size, + bigram_dim=args.bigram_dim, + xsa_last_n=args.xsa_last_n, + rope_dims=args.rope_dims, + ln_scale=args.ln_scale, + dtg=args.dtg_enabled, + ve_enabled=args.ve_enabled, + ve_dim=args.ve_dim, + ve_layers=args.ve_layers, + gated_attention=args.gated_attention, + value_residual=args.value_residual, + ).to(device).bfloat16() + # Banks stay FP32 (like CastedLinear weights), cast to BF16 in forward + base_model.qo_bank.data = base_model.qo_bank.data.float() + base_model.kv_bank.data = base_model.kv_bank.data.float() + base_model.mlp_up_bank.data = base_model.mlp_up_bank.data.float() + base_model.mlp_down_bank.data = base_model.mlp_down_bank.data.float() + for module in base_model.modules(): + if isinstance(module, CastedLinear): + module.float() + restore_low_dim_params_to_fp32(base_model) + # No DDP -- Parallel Muon handles bank grad communication via reduce-scatter, + # and non-bank grads are manually all-reduced before Adam steps. + compiled_model = torch.compile(base_model, dynamic=False, fullgraph=True) + model = compiled_model + + # Optimizer split: + # - 4 parameter banks -> Muon (batched Newton-Schulz) + # - token embedding -> Adam + # - scalars/control tensors -> Adam + # - bigram proj, mtp heads, VE proj -> Adam (small matrix params not worth banking) + matrix_params = [ + base_model.qo_bank, base_model.kv_bank, + base_model.mlp_up_bank, base_model.mlp_down_bank, + ] + block_named_params = list(base_model.blocks.named_parameters()) + scalar_params = [ + p + for name, p in block_named_params + if p.ndim < 2 or any(pattern in name for pattern in CONTROL_TENSOR_NAME_PATTERNS) + ] + if base_model.skip_weights.numel() > 0: + scalar_params.append(base_model.skip_weights) + scalar_params.append(base_model.smear.gate) + if base_model.bigram is not None: + scalar_params.append(base_model.bigram.scale) + token_lr = args.tied_embed_lr if args.tie_embeddings else args.embed_lr + tok_params = [{"params": [base_model.tok_emb.weight], "lr": token_lr, "base_lr": token_lr}] + if base_model.bigram is not None: + tok_params.append({"params": [base_model.bigram.embed.weight], "lr": token_lr, "base_lr": token_lr}) + if base_model.bigram.proj is not None: + scalar_params.append(base_model.bigram.proj.weight) + if base_model.ve_shared is not None: + tok_params.append({"params": [base_model.ve_shared.embed.weight], "lr": token_lr, "base_lr": token_lr}) + if base_model.ve_shared.proj is not None: + scalar_params.append(base_model.ve_shared.proj.weight) + scalar_params.append(base_model.ve_shared.scale) + for s in base_model.ve_layer_scales: + scalar_params.append(s) + optimizer_tok = torch.optim.AdamW( + tok_params, + betas=(args.beta1, args.beta2), + eps=args.adam_eps, + weight_decay=args.adam_wd, + fused=True, + ) + optimizer_muon = Muon( + matrix_params, + lr=args.matrix_lr, + momentum=args.muon_momentum, + backend_steps=args.muon_backend_steps, + weight_decay=args.muon_wd, + ) + for group in optimizer_muon.param_groups: + group["base_lr"] = args.matrix_lr + optimizer_scalar = torch.optim.AdamW( + [{"params": scalar_params, "lr": args.scalar_lr, "base_lr": args.scalar_lr}], + betas=(args.beta1, args.beta2), + eps=args.adam_eps, + weight_decay=args.adam_wd, + fused=True, + ) + # Non-bank params that need manual all-reduce (replicated across GPUs) + replicated_params = list(optimizer_tok.param_groups[0]["params"]) + for pg in optimizer_tok.param_groups[1:]: + replicated_params.extend(pg["params"]) + replicated_params.extend(scalar_params) + + optimizer_head = None + if base_model.lm_head is not None: + optimizer_head = torch.optim.Adam( + [{"params": [base_model.lm_head.weight], "lr": args.head_lr, "base_lr": args.head_lr}], + betas=(args.beta1, args.beta2), + eps=args.adam_eps, + fused=True, + ) + replicated_params.append(base_model.lm_head.weight) + optimizers: list[torch.optim.Optimizer] = [optimizer_tok, optimizer_muon, optimizer_scalar] + if optimizer_head is not None: + optimizers.append(optimizer_head) + n_params = sum(p.numel() for p in base_model.parameters()) + mtp_params = sum(p.numel() for p in base_model.mtp_heads.parameters()) + log0(f"model_params:{n_params}") + log0(f"mtp_num_heads:{args.mtp_num_heads} mtp_loss_weight:{args.mtp_loss_weight} mtp_params:{mtp_params}") + xsa_layers = [i for i, b in enumerate(base_model.blocks) if b.attn.use_xsa] + log0(f"XSA:last_{args.xsa_last_n} active_layers:{xsa_layers}") + log0(f"world_size:{world_size} grad_accum_steps:{grad_accum_steps}") + log0("sdp_backends:cudnn=False flash=True mem_efficient=False math=False") + log0(f"attention_mode:gqa num_heads:{args.num_heads} num_kv_heads:{args.num_kv_heads}") + log0( + f"tie_embeddings:{args.tie_embeddings} embed_lr:{token_lr} " + f"head_lr:{args.head_lr if base_model.lm_head is not None else 0.0} " + f"matrix_lr:{args.matrix_lr} scalar_lr:{args.scalar_lr}" + ) + log0( + f"train_batch_tokens:{args.train_batch_tokens} train_seq_len:{args.train_seq_len} " + f"iterations:{args.iterations} warmup_steps:{args.warmup_steps} " + f"max_wallclock_seconds:{args.max_wallclock_seconds:.3f}" + ) + log0(f"seed:{args.seed}") + train_loader = DistributedTokenLoader(args.train_files, rank, world_size, device) + def zero_grad_all() -> None: + for opt in optimizers: + opt.zero_grad(set_to_none=True) + max_wallclock_ms = 1000.0 * args.max_wallclock_seconds if args.max_wallclock_seconds > 0 else None + def lr_mul(step: int, elapsed_ms: float) -> float: + if args.warmdown_iters <= 0: + return 1.0 + if max_wallclock_ms is None: + warmdown_start = max(args.iterations - args.warmdown_iters, 0) + return max((args.iterations - step) / max(args.warmdown_iters, 1), 0.0) if warmdown_start <= step < args.iterations else 1.0 + step_ms = elapsed_ms / max(step, 1) + warmdown_ms = args.warmdown_iters * step_ms + remaining_ms = max(max_wallclock_ms - elapsed_ms, 0.0) + return remaining_ms / max(warmdown_ms, 1e-9) if remaining_ms <= warmdown_ms else 1.0 + if args.warmup_steps > 0: + initial_model_state = {name: tensor.detach().cpu().clone() for name, tensor in base_model.state_dict().items()} + initial_optimizer_states = [copy.deepcopy(opt.state_dict()) for opt in optimizers] + model.train() + for warmup_step in range(args.warmup_steps): + zero_grad_all() + for micro_step in range(grad_accum_steps): + x, y = train_loader.next_batch(args.train_batch_tokens, args.train_seq_len, grad_accum_steps) + with torch.autocast(device_type="cuda", dtype=torch.bfloat16, enabled=True): + warmup_loss = model(x, y) + (warmup_loss * grad_scale).backward() + # All-reduce all grads for warmup (simple, not optimized) + if distributed: + for p in base_model.parameters(): + if p.grad is not None: + dist.all_reduce(p.grad, op=dist.ReduceOp.AVG) + for opt in optimizers: + opt.step() + zero_grad_all() + if args.warmup_steps <= 20 or (warmup_step + 1) % 10 == 0 or warmup_step + 1 == args.warmup_steps: + log0(f"warmup_step:{warmup_step + 1}/{args.warmup_steps}") + base_model.load_state_dict(initial_model_state, strict=True) + for opt, state in zip(optimizers, initial_optimizer_states, strict=True): + opt.load_state_dict(state) + zero_grad_all() + train_loader = DistributedTokenLoader(args.train_files, rank, world_size, device) + swa_state: dict[str, Tensor] | None = None + swa_count = 0 + from collections import deque + lawa_queue: deque[dict[str, Tensor]] = deque(maxlen=args.lawa_k) + ema_state = {name: t.detach().float().clone() for name, t in base_model.state_dict().items()} + ema_decay = 0.997 + training_time_ms = 0.0 + stop_after_step: int | None = None + torch.cuda.synchronize() + t0 = time.perf_counter() + step = 0 + while True: + last_step = step == args.iterations or (stop_after_step is not None and step >= stop_after_step) + should_validate = last_step or (args.val_loss_every > 0 and step % args.val_loss_every == 0) + if should_validate: + torch.cuda.synchronize() + training_time_ms += 1000.0 * (time.perf_counter() - t0) + val_loss, val_bpb = eval_val( + args, + model, + rank, + world_size, + device, + grad_accum_steps, + val_tokens, + base_bytes_lut, + has_leading_space_lut, + is_boundary_token_lut, + ) + log0( + f"step:{step}/{args.iterations} val_loss:{val_loss:.4f} val_bpb:{val_bpb:.4f} " + f"train_time:{training_time_ms:.0f}ms step_avg:{training_time_ms / max(step, 1):.2f}ms" + ) + torch.cuda.synchronize() + t0 = time.perf_counter() + if last_step: + if stop_after_step is not None and step < args.iterations: + log0( + f"stopping_early: wallclock_cap train_time:{training_time_ms:.0f}ms " + f"step:{step}/{args.iterations}" + ) + break + elapsed_ms = training_time_ms + 1000.0 * (time.perf_counter() - t0) + scale = lr_mul(step, elapsed_ms) + if args.late_qat_threshold > 0 and scale < args.late_qat_threshold and not CastedLinear._qat_enabled: + CastedLinear._qat_enabled = True + log0(f"late_qat:enabled step:{step} scale:{scale:.4f}") + zero_grad_all() + train_loss = torch.zeros((), device=device) + for micro_step in range(grad_accum_steps): + x, y = train_loader.next_batch(args.train_batch_tokens, args.train_seq_len, grad_accum_steps) + with torch.autocast(device_type="cuda", dtype=torch.bfloat16, enabled=True): + loss = model(x, y) + train_loss += loss.detach() + (loss * grad_scale).backward() + train_loss /= grad_accum_steps + frac = min(step / args.muon_momentum_warmup_steps, 1.0) if args.muon_momentum_warmup_steps > 0 else 1.0 + muon_momentum = (1 - frac) * args.muon_momentum_warmup_start + frac * args.muon_momentum + for group in optimizer_muon.param_groups: + group["momentum"] = muon_momentum + for opt in optimizers: + for group in opt.param_groups: + group["lr"] = group["base_lr"] * scale + if args.grad_clip_norm > 0: + torch.nn.utils.clip_grad_norm_(base_model.parameters(), args.grad_clip_norm) + # === 3-phase overlapped optimizer step === + # Phase 1: Launch async reduce-scatter for banks (biggest first) + optimizer_muon.launch_reduce_scatters() + # Phase 2: All-reduce non-bank grads + step Adam (while bank RS is in-flight) + if distributed: + for p in replicated_params: + if p.grad is not None: + dist.all_reduce(p.grad, op=dist.ReduceOp.AVG) + optimizer_tok.step() + optimizer_scalar.step() + if optimizer_head is not None: + optimizer_head.step() + # Phase 3: Wait for RS, local NS5, all-gather (banks processed last) + optimizer_muon.step() + zero_grad_all() + # EMA update + with torch.no_grad(): + for name, t in base_model.state_dict().items(): + ema_state[name].mul_(ema_decay).add_(t.detach().float(), alpha=1.0 - ema_decay) + step += 1 + approx_training_time_ms = training_time_ms + 1000.0 * (time.perf_counter() - t0) + if args.swa_enabled and scale < 0.2 and step % args.swa_every == 0: + if swa_state is None: + swa_state = {name: t.detach().cpu().clone() for name, t in base_model.state_dict().items()} + swa_count = 1 + log0(f"swa:start step:{step}") + else: + for name, t in base_model.state_dict().items(): + swa_state[name] += t.detach().cpu() + swa_count += 1 + if args.lawa_enabled and step % args.lawa_freq == 0: + lawa_queue.append({name: t.detach().cpu().clone() for name, t in base_model.state_dict().items()}) + should_log_train = ( + args.train_log_every > 0 + and (step <= 10 or step % args.train_log_every == 0 or stop_after_step is not None) + ) + if should_log_train: + log0( + f"step:{step}/{args.iterations} train_loss:{train_loss.item():.4f} " + f"train_time:{approx_training_time_ms:.0f}ms step_avg:{approx_training_time_ms / step:.2f}ms" + ) + reached_cap = max_wallclock_ms is not None and approx_training_time_ms >= max_wallclock_ms + if distributed and max_wallclock_ms is not None: + reached_cap_tensor = torch.tensor(int(reached_cap), device=device) + dist.all_reduce(reached_cap_tensor, op=dist.ReduceOp.MAX) + reached_cap = bool(reached_cap_tensor.item()) + if stop_after_step is None and reached_cap: + stop_after_step = step + log0( + f"peak memory allocated: {torch.cuda.max_memory_allocated() // 1024 // 1024} MiB " + f"reserved: {torch.cuda.max_memory_reserved() // 1024 // 1024} MiB" + ) + # Apply weight averaging + if args.lawa_enabled and len(lawa_queue) > 1: + log0(f"lawa:applying LAWA averaging k={len(lawa_queue)}") + current_state = base_model.state_dict() + avg_state = {name: torch.zeros(t.shape, dtype=torch.float32, device='cpu') for name, t in current_state.items()} + for snap in lawa_queue: + for name in avg_state: + avg_state[name] += snap[name].float() + for name in avg_state: + avg_state[name] /= len(lawa_queue) + avg_state[name] = avg_state[name].to(dtype=current_state[name].dtype) + base_model.load_state_dict(avg_state, strict=True) + else: + log0("ema:applying EMA weights") + current_state = base_model.state_dict() + avg_state = {name: t.to(dtype=current_state[name].dtype) for name, t in ema_state.items()} + base_model.load_state_dict(avg_state, strict=True) + torch.cuda.synchronize() + t_diag = time.perf_counter() + diag_val_loss, diag_val_bpb = eval_val( + args, compiled_model, rank, world_size, device, grad_accum_steps, + val_tokens, base_bytes_lut, has_leading_space_lut, is_boundary_token_lut, + ) + torch.cuda.synchronize() + log0( + f"DIAGNOSTIC post_ema val_loss:{diag_val_loss:.4f} val_bpb:{diag_val_bpb:.4f} " + f"eval_time:{1000.0 * (time.perf_counter() - t_diag):.0f}ms" + ) + full_state_dict = base_model.state_dict() + export_sd = {k: v for k, v in full_state_dict.items() if "mtp_heads" not in k} + excluded_mtp = sum(int(t.numel()) for k, t in full_state_dict.items() if "mtp_heads" in k) + if excluded_mtp > 0: + log0(f"export_excluding_mtp_params:{excluded_mtp}") + if master_process: + torch.save(export_sd, "final_model.pt") + model_bytes = os.path.getsize("final_model.pt") + code_bytes = len(code.encode("utf-8")) + log0(f"Serialized model: {model_bytes} bytes") + log0(f"Code size: {code_bytes} bytes") + # Unbank 3D tensors into individual 2D tensors for quantization + sd_cpu = {k: v.detach().cpu() for k, v in export_sd.items()} + unbanked_sd = _unbank_state_dict(sd_cpu, args.num_layers) + # Full GPTQ: collect Hessians via a temporary non-banked model + log0(f"gptq:building non-banked model for Hessian collection...") + hessian_model = _HessianGPT( + vocab_size=args.vocab_size, num_layers=args.num_layers, model_dim=args.model_dim, + num_heads=args.num_heads, num_kv_heads=args.num_kv_heads, mlp_mult=args.mlp_mult, + tie_embeddings=args.tie_embeddings, logit_softcap=args.logit_softcap, + rope_base=args.rope_base, qk_gain_init=args.qk_gain_init, + bigram_vocab_size=args.bigram_vocab_size, bigram_dim=args.bigram_dim, + xsa_last_n=args.xsa_last_n, rope_dims=args.rope_dims, ln_scale=args.ln_scale, + ve_enabled=args.ve_enabled, ve_dim=args.ve_dim, ve_layers=args.ve_layers, + ).to(device).bfloat16() + for m in hessian_model.modules(): + if isinstance(m, CastedLinear): + m.float() + restore_low_dim_params_to_fp32(hessian_model) + # Load unbanked weights into the non-banked model + hessian_model.load_state_dict( + {k: v.to(device) for k, v in unbanked_sd.items() if k in hessian_model.state_dict()}, + strict=False, + ) + # Autoregressive self-generated calibration (no external data) + log0("gptq:generating autoregressive calibration data (64 seqs x 2048 tokens, temp=0.8)...") + base_model.load_state_dict(export_sd, strict=False) + t_gen = time.perf_counter() + ar_tokens = generate_autoregressive_calib( + base_model, device, num_seqs=64, seq_len=args.train_seq_len, + vocab_size=args.vocab_size, temperature=0.8, batch_size=8, seed=args.seed, + ) + log0(f"gptq:generated {len(ar_tokens)} sequences in {time.perf_counter()-t_gen:.1f}s") + log0("gptq:collecting hessians from autoregressive data...") + hessians = collect_hessians_from_tokens(hessian_model, ar_tokens, device) + log0(f"gptq:collected hessians for {len(hessians)} layers (AR self-gen)") + del ar_tokens + del hessian_model + torch.cuda.empty_cache() + quant_result, quant_meta = mixed_quantize_int6(unbanked_sd, {"mlp", "attn"}, hessians=hessians) + # NOVEL: Selective ±1 pruning by reconstruction error + # Sort ±1 quantized values by their reconstruction error (scale²), + # prune least-impactful first until artifact fits target size. + target_mb = float(os.environ.get("TARGET_MB", "15.9")) + code_bytes_est = len(code.encode("utf-8")) + ones_info = [] # (tensor_key, flat_idx, error) + for name, info in quant_meta.items(): + if not (isinstance(info, dict) and info.get("type") == "int6"): continue + qk, sk = name + ".q", name + ".scale" + if qk not in quant_result or sk not in quant_result: continue + q, s = quant_result[qk], quant_result[sk] + if s.ndim > 0: + ones_mask = (q.abs() == 1) + if ones_mask.any(): + row_idx = torch.arange(q.shape[0]).unsqueeze(1).expand_as(q)[ones_mask] + flat_idx = torch.arange(q.numel()).reshape(q.shape)[ones_mask] + errors = s.float()[row_idx].pow(2) + for fi, err in zip(flat_idx.tolist(), errors.tolist()): + ones_info.append((qk, fi, err)) + if ones_info: + ones_info.sort(key=lambda x: x[2]) + def _try_prune(n): + tmp = {k: v.clone() for k, v in quant_result.items()} + for i in range(min(n, len(ones_info))): + tmp[ones_info[i][0]].view(-1)[ones_info[i][1]] = 0 + buf = io.BytesIO(); torch.save({"w": tmp, "m": quant_meta}, buf) + return len(lzma.compress(buf.getvalue(), preset=9)) + code_bytes_est, tmp + no_sz, _ = _try_prune(0) + target_bytes = int(target_mb * 1024 * 1024) + log0(f"selective_prune: {len(ones_info)} ±1 candidates, unpruned={no_sz/(1024*1024):.2f}MB target={target_mb}MB") + if no_sz <= target_bytes: + log0("selective_prune: already fits, no pruning needed") + else: + full_sz, _ = _try_prune(len(ones_info)) + log0(f"selective_prune: full ±1 prune={full_sz/(1024*1024):.2f}MB") + if full_sz > target_bytes: + log0("selective_prune: even full prune not enough, applying all") + _, quant_result = _try_prune(len(ones_info)) + else: + lo, hi = 0, len(ones_info) + while lo < hi: + mid = (lo + hi) // 2 + sz, _ = _try_prune(mid) + if sz <= target_bytes: hi = mid + else: lo = mid + 1 + log0(f"selective_prune: pruning {lo}/{len(ones_info)} ±1 values ({100*lo/len(ones_info):.1f}%) to fit {target_mb}MB") + _, quant_result = _try_prune(lo) + quant_buf = io.BytesIO() + torch.save({"w": quant_result, "m": quant_meta}, quant_buf) + quant_raw = quant_buf.getvalue() + quant_blob = lzma.compress(quant_raw, preset=9) + if master_process: + with open("final_model.int6.ptz", "wb") as f: + f.write(quant_blob) + quant_file_bytes = len(quant_blob) + code_bytes = len(code.encode("utf-8")) + log0(f"Serialized model int6+lzma: {quant_file_bytes} bytes") + log0(f"Total submission size int6+lzma: {quant_file_bytes + code_bytes} bytes") + if distributed: + dist.barrier() + with open("final_model.int6.ptz", "rb") as f: + quant_blob_disk = f.read() + quant_state = torch.load( + io.BytesIO(lzma.decompress(quant_blob_disk)), + map_location="cpu", + ) + deq_unbanked = dequantize_mixed_int6(quant_state["w"], quant_state["m"], unbanked_sd) + # Re-bank the dequantized tensors + deq_state = _rebank_state_dict(deq_unbanked, args.num_layers, sd_cpu) + eval_model = GPT( + vocab_size=args.vocab_size, num_layers=args.num_layers, model_dim=args.model_dim, + num_heads=args.num_heads, num_kv_heads=args.num_kv_heads, mlp_mult=args.mlp_mult, + tie_embeddings=args.tie_embeddings, tied_embed_init_std=args.tied_embed_init_std, + logit_softcap=args.logit_softcap, rope_base=args.rope_base, qk_gain_init=args.qk_gain_init, + mtp_num_heads=0, mtp_loss_weight=0.0, + bigram_vocab_size=args.bigram_vocab_size, bigram_dim=args.bigram_dim, + xsa_last_n=args.xsa_last_n, + rope_dims=args.rope_dims, ln_scale=args.ln_scale, dtg=args.dtg_enabled, + ve_enabled=args.ve_enabled, ve_dim=args.ve_dim, ve_layers=args.ve_layers, + gated_attention=args.gated_attention, value_residual=args.value_residual, + ).to(device).bfloat16() + eval_model.qo_bank.data = eval_model.qo_bank.data.float() + eval_model.kv_bank.data = eval_model.kv_bank.data.float() + eval_model.mlp_up_bank.data = eval_model.mlp_up_bank.data.float() + eval_model.mlp_down_bank.data = eval_model.mlp_down_bank.data.float() + for m in eval_model.modules(): + if isinstance(m, CastedLinear): + m.float() + restore_low_dim_params_to_fp32(eval_model) + eval_model.load_state_dict(deq_state, strict=True) + compiled_eval = torch.compile(eval_model, dynamic=False, fullgraph=True) + torch.cuda.synchronize() + t_qeval = time.perf_counter() + q_val_loss, q_val_bpb = eval_val( + args, compiled_eval, rank, world_size, device, grad_accum_steps, + val_tokens, base_bytes_lut, has_leading_space_lut, is_boundary_token_lut, + eval_seq_len=effective_eval_seq_len, + ) + torch.cuda.synchronize() + log0( + f"final_int6_roundtrip val_loss:{q_val_loss:.4f} val_bpb:{q_val_bpb:.4f} " + f"eval_time:{1000.0 * (time.perf_counter() - t_qeval):.0f}ms" + ) + log0(f"final_int6_roundtrip_exact val_loss:{q_val_loss:.8f} val_bpb:{q_val_bpb:.8f}") + sw_seq_len = effective_eval_seq_len + if args.eval_stride > 0 and args.eval_stride < sw_seq_len: + torch.cuda.synchronize() + t_slide = time.perf_counter() + sw_val_loss, sw_val_bpb = eval_val_sliding( + args, eval_model, rank, world_size, device, + val_tokens, base_bytes_lut, has_leading_space_lut, is_boundary_token_lut, + stride=args.eval_stride, + eval_seq_len=sw_seq_len, + ) + torch.cuda.synchronize() + log0( + f"final_int6_sliding_window val_loss:{sw_val_loss:.4f} val_bpb:{sw_val_bpb:.4f} " + f"stride:{args.eval_stride} eval_time:{1000.0 * (time.perf_counter() - t_slide):.0f}ms" + ) + log0(f"final_int6_sliding_window_exact val_loss:{sw_val_loss:.8f} val_bpb:{sw_val_bpb:.8f}") + log0(f"final_int8_zlib_roundtrip_exact val_loss:{sw_val_loss:.8f} val_bpb:{sw_val_bpb:.8f}") + if args.eval_stride != 64 and 64 < sw_seq_len: + torch.cuda.synchronize() + t_slide64 = time.perf_counter() + sw64_val_loss, sw64_val_bpb = eval_val_sliding( + args, eval_model, rank, world_size, device, + val_tokens, base_bytes_lut, has_leading_space_lut, is_boundary_token_lut, + stride=64, + eval_seq_len=sw_seq_len, + ) + torch.cuda.synchronize() + log0( + f"final_int6_sliding_window_s64 val_loss:{sw64_val_loss:.4f} val_bpb:{sw64_val_bpb:.4f} " + f"stride:64 eval_time:{1000.0 * (time.perf_counter() - t_slide64):.0f}ms" + ) + log0(f"final_int6_sliding_window_s64_exact val_loss:{sw64_val_loss:.8f} val_bpb:{sw64_val_bpb:.8f}") + log0(f"final_int8_zlib_roundtrip_exact val_loss:{sw64_val_loss:.8f} val_bpb:{sw64_val_bpb:.8f}") + if distributed: + dist.destroy_process_group() +if __name__ == "__main__": + main()