v0.5.0

What's Changed

• init mindnlp 0.5.0 with mindtorch by @lvyufeng in #2016
• fix bugs on transformers 4.51.1 by @lvyufeng in #2017
• 【开源实习】Bartpho模型微调 by @SeasonMay in #2030
• update qwen jit inference by @ldqLDQ in #2041
• 【开源实习】Barthez模型微调 by @SeasonMay in #2032
• use new core to adapt huggingface library by @lvyufeng in #2056
• use lazymodule to load transformers submodules by @lvyufeng in #2057
• fix some bugs for transformers ut(bert) by @lvyufeng in #2058
• use huggingface transformers official ut to test by @lvyufeng in #2059
• fix proxy error and pass more hf ut by @lvyufeng in #2060
• fix all bert ut and skip useless ut by @lvyufeng in #2061
• pass all llama ut on Ascend by @lvyufeng in #2062
• fix qwen series bug by @lvyufeng in #2063
• fix lora training by @lvyufeng in #2064
• fix albert by @lvyufeng in #2065
• fix autoformer by @lvyufeng in #2066
• fix a class models ut by @lvyufeng in #2067
• fix b class ut by @lvyufeng in #2068
• fix c class ut by @lvyufeng in #2069
• fix d class ut by @lvyufeng in #2070
• fix e class ut by @lvyufeng in #2071
• fix e class by @lvyufeng in #2072
• fix g class ut by @lvyufeng in #2073
• fix h class ut by @lvyufeng in #2074
• fix i class ut by @lvyufeng in #2075
• fix l class ut by @lvyufeng in #2076
• fix m class ut by @lvyufeng in #2077
• fix n class ut by @lvyufeng in #2078
• fix o class ut by @lvyufeng in #2079
• fix diffuser models ut by @lvyufeng in #2080
• fix diffusers unet ut by @lvyufeng in #2081
• fix transformers p class ut by @lvyufeng in #2082
• fix transformers r class ut by @lvyufeng in #2083
• fix diffuers pipelines c class ut by @lvyufeng in #2084
• fix diffusers pipelines d class ut by @lvyufeng in #2085
• fix diffusers pipeline h class ut by @lvyufeng in #2086
• fix diffusers pipelines k class ut by @lvyufeng in #2087
• fix transformers u class ut by @lvyufeng in #2088
• fix diffusers pipelines l class ut by @lvyufeng in #2089
• fix transformers v class ut by @lvyufeng in #2090
• fix diffusers omni ut by @lvyufeng in #2091
• fix transformers w class ut and diffusers omni by @lvyufeng in #2092
• fix torchvision 0.16 import error by @lvyufeng in #2093
• update readme and fix diffuser pipelines p class ut by @lvyufeng in #2094
• fix transformers x class ut by @lvyufeng in #2095
• fix diffusers pipelines s class ut by @lvyufeng in #2096
• fix diffusers pipelines wan ut by @lvyufeng in #2097
• fix diffusers lora/schedulers ut by @lvyufeng in #2098
• transformers.pipeline support ms_dtype/mindspore_dtype by @lvyufeng in #2099
• diffusers.DiffusionPipeline support ms_dtype/mindspore_dtype by @lvyufeng in #2100
• fix data folder ignore by @lvyufeng in #2103
• fix ms2.6 & bert on gpu by @lvyufeng in #2104
• fix qwen3/qwen3_moe/llama on GPU by @lvyufeng in #2105
• auto replace cuda with npu when use Ascend device by @lvyufeng in #2106
• fix normal memory issue on GPU by @lvyufeng in #2107
• fix diffusers autoencoder on GPU by @lvyufeng in #2108
• fix core.where on GPU by @lvyufeng in #2109
• fix ops.where on GPU by @lvyufeng in #2111
• speedup sdpa_mask for mindspore by @lvyufeng in #2112
• speedup sdpa_mask for mindspore by @lvyufeng in #2113
• add diffusers notebooks by @lvyufeng in #2114
• 【开源实习】blip模型微调 by @Alemax067 in #1896
• 【开源实习】bert模型微调 by @Alemax067 in #1897
• 【开源实习】TrOCR模型应用开发 by @jiuniZZzz in #1908
• T5模型应用开发 by @wyx-lab in #1920
• 【开源实习】ConvNeXT模型应用开发 by @AM-SuSh in #1946
• 【开源实习】Perceiver IO模型应用开发 by @AM-SuSh in #1947
• 【开源实习】LayoutXLM模型应用开发 by @Majianb in #1970
• 开源实习-LayoutLM模型应用开发-金逸 by @MQN-80 in #1971
• 【开源实习】bigbird_pegasus模型微调 by @outbreak-sen in #1972
• 【开源实习】chatglm-4模型微调 by @shee333 in #1996
• 【开源实习】GIT模型应用开发&&LayoutLMv3模型应用开发 by @wujiu0411 in #2027
• 【开源实习】CANNIE模型应用开发 by @ToAdventure in #2044
• 【开源实习】altclip模型微调 by @aorangehc in #2049
• add diffusers doc notebooks by @lvyufeng in #2115
• fix bert modeling ut on ORAGNE PI by @lvyufeng in #2116
• fix transformers ut for 4.54.1 by @lvyufeng in #2117
• use gather_nd to implement gather for Orange Pi by @lvyufeng in #2118
• fix apis for Orange Pi by @lvyufeng in #2119
• fix qwen2_vl in orange pi by @lvyufeng in #2122
• support multiprocess inference on 910A by @lvyufeng in #2124
• fix gpt-oss precision and add examples by @lvyufeng in #2125
• fix merge bugs by @lvyufeng in #2126
• support ascend310b prims by @lvyufeng in #2127
• update gpt-oss app by @lvyufeng in #2128
• fix bark ut and some numpy prims by @lvyufeng in #2130
• fix peft memory cost issue by @lvyufeng in #2133
• fix layer_norm with bias=None by @lvyufeng in #2134
• fix qwen3/qwen3 moe by @lvyufeng in #2135
• add core.testing for torch ut by @lvyufeng in #2136
• fix qwen2.5 lora finetune on 910A by @lvyufeng in #2137
• fix qwen2.5 finetune precision with sdpa by @lvyufeng in #2138
• fix import error with accelearte==1.10 by @lvyufeng in #2139
• fix bfloat16 not found for numpy op by @lvyufeng in #2141
• sdpa use flash attention for 910B by @lvyufeng in #2142
• fix flash attention on 910B when is_causal=True by @lvyufeng in #2143
• fix some api bugs for new dispatcher by @lvyufeng in #2144
• pass bert ut on Orange PI by @lvyufeng in #2145
• fix core.is_complex by @lvyufeng in #2146
• fix apis for c class by @lvyufeng in #2149
• fix inplace copy on cpu by @lvyufeng in #2150
• fix apis for d class by @lvyufeng in #2151
• fix apis for e class by @lvyufeng in #2152
• fix apis for f class by @lvyufeng in #2153
• fix apis for g class by @lvyufeng in #2155
• fix apis for h-k class by @lvyufeng in #2156
• fix apis for l-n class by @lvyufeng in #2157
• fix apis for o-r class by @lvyufeng in #2158
• fix apis for s-z class by @lvyufeng in #2159
• upgrade readme and version by @lvyufeng in #2160
• fix janus demo with device dispatch by @lvyufeng in #2161
• refactor prim op for different backend by @lvyufeng in #2164
• fix tokenizer not support return_tensor='ms' by @lvyufeng in #2165
• fix a class on GPU by @lvyufeng in #2166
• fix b class on GPU by @lvyufeng in #2167
• move mindnlp.core to mindtorch by @lvyufeng in #2168
• add distributed.run to implement device mesh by @lvyufeng in #2169
• try migrate dtensor by @lvyufeng in #2170
• fix x-z class on GPU by @lvyufeng in #2171
• fix u-w class on GPU by @lvyufeng in #2172
• fix torch.version by @lvyufeng in #2173
• fix empty_instance for ms2.7 by @lvyufeng in #2174
• fix r-t class on GPU by @lvyufeng in #2175
• support env var CPU_USE_NUMPY by @lvyufeng in #2176
• fix o-q class on GPU by @lvyufeng in #2177
• fix l-n class on GPU by @lvyufeng in #2178
• fix e-g class on GPU by @lvyufeng in #2180
• fix a-d class on GPU by @lvyufeng in #2181
• fix ms2.7 on GPU by @lvyufeng in #2182
• remove legacy code and fix pylint by @lvyufeng in #2183
• add requirements by @lvyufeng in #2184
• fix qwen_vl and load_pretrained patch by @lvyufeng in #2190
• add triu_indices by @lvyufeng in #2191
• fix bert on OrangePi by @lvyufeng in #2193
• add:open_r1 by @ResDream in #2179
• fix stable diffusion on npu by @lvyufeng in #2194
• fix empty with ms2.6 by @lvyufeng in #2195
• support sentence transformers by @lvyufeng in #2198
• update version by @lvyufeng in #2199
• fix conv2d on OrangePi by @lvyufeng in #2200
• add fmod_tensor by @lvyufeng in #2202
• add paddleocr-vl example by @lvyufeng in #2203
• fix a class models on OrangePi by @lvyufeng in #2205
• add new rag method by @moyu026 in #2204
• fix b class ut on OrangePi by @lvyufeng in #2206
• fix bi* models on OrangePi by @lvyufeng in #2207
• support deepseek-ocr by @lvyufeng in #2208
• fix b class models on OrangePi by @lvyufeng in #2210
• fix q class models on OrangePi by @lvyufeng in #2211
• add mindtorch.lerp by @lvyufeng in #2212
• fix c class models on OrangePi by @lvyufeng in #2213
• fix d class models on OrangePi by @lvyufeng in #2214
• fix e class on OrangePi by @lvyufeng in #2215
• fix f class on OrangePi by @lvyufeng in #2216
• try torch_dispatch method by @lvyufeng in #2217
• fix g class on OrangePi by @lvyufeng in #2219
• fix h class model on OrangePi by @lvyufeng in #2220
• fix i-l class on OrangePi by @lvyufeng in #2222
• fix m,n class on OrangePi by @lvyufeng in #2223
• fix o-s class on OrangePi by @lvyufeng in #2227
• fix t-z class on OrangePi by @lvyufeng in #2228
• update readme and git pipeline by @lvyufeng in #2229

New Contributors

• @SeasonMay made their first contribution in #2030
• @jiuniZZzz made their first contribution in #1908
• @wyx-lab made their first contribution in #1920
• @AM-SuSh made their first contribution in #1946
• @Majianb made their first contribution in #1970
• @MQN-80 made their first contribution in #1971
• @shee333 made their first contribution in #1996
• @wujiu0411 made their first contribution in #2027
• @ToAdventure made their first contribution in #2044
• @aorangehc made their first contribution in #2049

Full Changelog: v0.4.1...v0.5.0