-
When fine-tuning LLM using train.csv, does the sample require the full template which includes the **bos and eos?**
For example, if the model bos_token is ``````, do I need to include it into the t…
-
Seems like I cannot resolve the packages correctly with 0.20:
```
swift package update …
-
Request:
```
import requests
import json
url = "http://localhost:8000/v1/chat/completions"
headers = {"Content-Type": "application/json"}
data = {
"model": "mlx-community/Phi-3-mini-128…
-
Primarily to be used with Phi-2
-
use modelscope jupiter:
!pip install modelscope
import os
os.environ['VLLM_USE_MODELSCOPE'] = 'True'
!pip install vllm
!pip install --upgrade transformers==4.37.2
-----
ms-swift 1.5.1 re…
-
Hello, I'm trying to run inference of a qwen-vl-chat model that I've finetuned.
I'm using a 4xV100 cluster, how do I let the inference have access to all 4 gpus? How do I change `os.environ['CUDA_VI…
-
另外，请问`public/icon`下的：
android-chrome-192x192.png apple-touch-icon.png favicon-32x32.png
android-chrome-512x512.png favicon-16x16.png
都是干什么的，谢谢
-
Models are taking really good time to download. If the WiFI is OFF for a second it gets struck - it would be great if the example has two more options - Delete and Retry.
-
There seems to be Gemma support (https://github.com/ml-explore/mlx-swift-examples/blob/main/Libraries/LLM/Gemma.swift), however, the tokenizer library (when using `llm-tool`) throws a `unsupportedToke…
rudro updated
8 months ago
-
@shelby3 [wrote](https://github.com/keean/traitscript/issues/2#issuecomment-248102324):
> Also, the compiler can use a much simpler form of path analysis (independent of run-time state) to do that pe…