-
Hi, I was wondering if it would be possible to integrate chat arena with something like gymnasium or pettingzoo (multi agent version of gym). Would be very interesting to see LLMs be used both for lan…
-
I am using TRL and PEFT to finetune models. You can see the code I am using [here](https://github.com/mallorbc/Finetune_LLMs/blob/main/finetuning_repo/trl_finetune.py)
When finetuning Llama 7B, I …
-
### Describe the bug
When attempting to install the newest (and older) versions of scikit-learn-1.6.dev0 using regular pip install, the installation fails due to a missing version.py file. This iss…
-
OK I have one more request I'm hoping you can implement for this wonderful app.
The way I use the app, and I imagine others would too, is I am testing various models & verifying the performance to…
Hisma updated
8 months ago
-
Thanks for the great project.
I gave a try and compared with hf's offline inference speed on 100 alpaca examples. The hardware I used is a single v100-40G GPU. Here is my script for vLLM:
```
sam…
-
#### **Motivation**:
With the advent of powerful language models such as ChatGPT, there exists a potential to harness the power of these models in assisting developers in real-time as they interact…
-
https://github.com/intel-analytics/BigDL/tree/7a1a9edca7ebc1c02e155828f8475eb51cb8c06b/python/llm/example/GPU/PyTorch-Models/Model/qwen-vl
bigdl2.5.0b20240311
CPU memory:32G
ADL+A770
因为32G内存无法加载…
-
知乎每日精选 2024-06-23 更新
-
i was looking at the feature and wrote a lot of people on it.. i think they know its the bets way you speak to the object model not to the keyboartd adn mouse. and your more traditoinal NLP can it …
-
from .ds.ds1000 import DS1000Dataset
*** ModuleNotFoundError: No module named 'lm_eval.ds'