Tamminhdiep97

> Such as：mobilenet-v2, I trained 40 epochs,The loss dropped very slowly, only from 30 to 25. I Tried on EfficientNet-B4, loss also dropped so so so slow.

much appreciated

Im also using V100 Currently i modified the loading model as follow, notice the torch type ``` model = LlamaForCausalLM.from_pretrained( base_model, load_in_8bit=False, torch_dtype=torch.float32, llm_int8_skip_modules=FULL_FINETUNE_MODULES, device_map=device_map, cache_dir='../huggingface' ) ``` and add...

you should increase **micro_batch_size** param @Tungsong

python 2, theano 0.8.0

> @Solomin0, in your main change your HttpClient to the following: > > ```python > chroma_client = chromadb.HttpClient(host="chromaDB", port = 8000, settings=Settings(allow_reset=True, anonymized_telemetry=False)) > ``` for me, this solution work,...

hi, @cin-kay here are my docker-compose file: ``` version: '3.5' services: chatbot_engine: container_name: chatbot_engine # restart: always build: context: ../chatbot_util dockerfile: ../chatbot_util/Dockerfile image: chatbot_engine:dev environment: - PYTHONUNBUFFERED=1 - TRANSFORMERS_CACHE=./cache/ volumes:...

> Hi, I am having a similar issue, despite having tried the different solutions proposed here I still get the following error: > > > Traceback (most recent call last):...