v0.4.0-rc.2

Summary of Changes

• Support for LoRA tuning for llama3 and granite (with GPTBigCode) architectures
• Various dependencies versions adjustment

What's Changed

• remove merge model for lora tuned adapters by @anhuong in #197
• Add test coverage by @tedhtchang in #171
• Install Acceleration Framework into Training Script by @fabianlim in #157
• deps: limit dependency ranges by @anhuong in #54
• Delete dependabot.yml by @tedhtchang in #207
• add dependabot.yml by @tedhtchang in #208
• Fix additional callbacks by @VassilisVassiliadis in #199
• Update trl by @alex-jw-brooks in #213
• deps: cap transformers at 4.40.2 by @anhuong in #218

Full Changelog: v0.3.0...v0.4.0-rc.2