KeyError: 'llava_llama'

[rohithbojja]

(llavalora) PS F:\finetune_LLaVA> python llava\eval\run_llava.py --model-path C:\Users\rohit\Downloads\Compressed\llama-2-7b-chat-task-qlora_5\workspace\LLaVA\model-lora\llama-2-7b-chat-task-qlora --model-base C:\models\llava-v1.5-7b --image-file "C:\datasets\evqa-rad\images\34bc8f62-7dc1-4491-ab51-5b521f2acfe5.jpg" --query "describe this image" bin C:\ProgramData\miniconda3\envs\llavalora\lib\site-packages\bitsandbytes\libbitsandbytes_cuda118.dll Traceback (most recent call last): File "F:\finetune_LLaVA\llava\eval\run_llava.py", line 157, in eval_model(args) File "F:\finetune_LLaVA\llava\eval\run_llava.py", line 56, in eval_model tokenizer, model, image_processor, context_len = load_pretrained_model( File "C:\ProgramData\miniconda3\envs\llavalora\lib\site-packages\llava\model\builder.py", line 50, in load_pretrained_model lora_cfg_pretrained = AutoConfig.from_pretrained(model_path) File "C:\ProgramData\miniconda3\envs\llavalora\lib\site-packages\transformers\models\auto\configuration_auto.py", line 998, in from_pretrained config_class = CONFIG_MAPPING[config_dict["model_type"]] File "C:\ProgramData\miniconda3\envs\llavalora\lib\site-packages\transformers\models\auto\configuration_auto.py", line 710, in getitem raise KeyError(key) KeyError: 'llava_llama'

[awzhgw]

i get same error . how toresolve it ?

[Celtic-sf]

I solved it by installing latest llava and transformers

[dhruv10xd]

I am facing the same error when i try to load my custom finetuned 7b lora model. Were you able to resolve it?