search

2U1 / Phi3-Vision-Finetune

An open-source implementaion for fine-tuning Phi3-Vision and Phi3.5-Vision by Microsoft.
Apache License 2.0
43 stars 7 forks source link

TypeError: Accelerator.save() got an unexpected keyword argument 'max_shard_size' #3

Closed tusharraskar closed 2 months ago

tusharraskar commented 2 months ago

bash scripts/merge_lora.sh Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained. Loading checkpoint shards: 100%|███| 2/2 [01:02<00:00, 31.25s/it] Traceback (most recent call last): File "/home/ubuntu/SB_Tushar/lora_phi3/Phi3-Vision-ft/src/merge_lora_weights.py", line 31, in merge_lora(args) File "/home/ubuntu/SB_Tushar/lora_phi3/Phi3-Vision-ft/src/merge_lora_weights.py", line 12, in merge_lora accel.save(model, args.save_model_path, max_shard_size = '5GB') TypeError: Accelerator.save() got an unexpected keyword argument 'max_shard_size'

2U1 commented 2 months ago

@tusharraskar I've updated the code with accel.save_model, and it wasn't updated proprely. Sorry for the inconvinience. You could change the function as I mentioned.

https://github.com/2U1/Phi3-Vision-ft/blob/e90cf1c4c74e895b7f80b46c8a05ae9fa27bc5d5/src/merge_lora_weights.py#L12