search

clcarwin / alpaca-weight

Train llama with lora on one 4090 and merge weight of lora to work as stanford alpaca.
MIT License
50 stars 4 forks source link

error while saving model #2

Open ChrisXULC opened 1 year ago

ChrisXULC commented 1 year ago

/root/anaconda3/envs/lmflow/lib/python3.9/site-packages/bitsandbytes/cuda_setup/main.py:136: UserWarning: /root/anaconda3/envs/lmflow did not contain libcudart.so as expected! Searching further paths... warn(msg) CUDA SETUP: CUDA runtime path found: /usr/local/cuda-11.4/lib64/libcudart.so CUDA SETUP: Highest compute capability among GPUs detected: 7.0 CUDA SETUP: Detected CUDA version 114 /root/anaconda3/envs/lmflow/lib/python3.9/site-packages/bitsandbytes/cuda_setup/main.py:136: UserWarning: WARNING: Compute capability < 7.5 detected! Only slow 8-bit matmul is supported for your GPU! warn(msg) CUDA SETUP: Loading binary /root/anaconda3/envs/lmflow/lib/python3.9/site-packages/bitsandbytes/libbitsandbytes_cuda114_nocublaslt.so... Loading checkpoint shards: 100%|██████████████████████████████████████████████| 3/3 [00:20<00:00, 6.96s/it]

Saving alpaca model to ./ Traceback (most recent call last): File "/home/LMFlow/cn_llama/packages/new/alpaca-weight/merge_step2_patch2alpaca.py", line 32, in trainer = Seq2SeqTrainer(model=model) File "/root/anaconda3/envs/lmflow/lib/python3.9/site-packages/transformers/trainer_seq2seq.py", line 72, in init if self.args.generation_config is not None: AttributeError: 'TrainingArguments' object has no attribute 'generation_config'

tomasmcz commented 1 year ago

Same thing happened to me.

bruvduroiu commented 1 year ago

Trying to address this here: https://github.com/clcarwin/alpaca-weight/pull/3

clcarwin commented 1 year ago

When writing the code, the LLaMA is not in the main branch of transformers. The pull request of bruvduroiu has been merged. Feel free to try and thanks. @bruvduroiu @tomasmcz @ChrisXULC