yuqiao9
yuqiao9
 Is it caused by a configuration error on my side?
> Hey, thanks for reporting this. > > The reason here is likely: ollama is using a 4-bit quantized model. However exo is using the fp16 unquantized model. > >...
I noticed there was an error when uploading the file. Could you please tell me what might be causing this
> 您能否仔细检查一下： > > * 您使用的是 OpenAI LLM 并设置了 GRAPHRAG_API_KEY 环境变量。 > * 您正在使用自定义模型，并确保那里的URL和设置正确。`settings.yaml.example` > > 如果您将 GraphRAG 与 Kotaemon Docker 版本一起使用，则可能需要更改主机名才能与主机上的 Ollama 或其他服务正确通信。查看 https://stackoverflow.com/questions/31324981/how-to-access-host-port-from-docker-container I believe my path is...
I downloaded the model completely from huggingface, and I don’t have folders like “snapshots” or the like. How do I get Exo to use the untreated model?Now I used the...
The current situation is that I have already used the --download-quick-check parameter, but the logs still show that it is downloading from Hugging Face because the local files cannot be...
Handling chat completions request from 192.168.8.3: {'model': 'llama-3.1-70b', 'messages': [{'role': 'user', 'content': '111'}], 'stream': True} Checking if local path exists to load tokenizer from local local_path=None Local check for local_path=None...
I’ve now changed the local_path in tokenizer.py from "/root/.cache/huggingface/hub/models--NousResearch--Meta-Llama-3.1-70B", return AutoTokenizer.from_pretrained("/root/.cache/huggingface/hub/models--NousResearch--Meta-Llama-3.1-70B"),and the error is as follows: Handling chat completions request from 192.168.8.3: {'model': 'llama-3.1-70b', 'messages': [{'role': 'user', 'content': 'sdsd'}], 'stream':...