SimFG comments

Results 402 comments of


                                            SimFG

Support the GPTCache to reduce the model Response Time

@Ph0rk0z have something updated? :)

[Feature]:Support audio embeddings

Have added a basic embedding, so we will close the issue. If there are more complex needs, we will open a new issue

The MPS backend is supported on MacOS 12.3+.Current OS version can be queried using `sw_vers`

@silasdao you can try add `--cpu` param when starting the server

The MPS backend is supported on MacOS 12.3+.Current OS version can be queried using `sw_vers`

@FrankDMartinez you can try replace `mps` to `cpu`

The MPS backend is supported on MacOS 12.3+.Current OS version can be queried using `sw_vers`

change all, you can try it ![image](https://user-images.githubusercontent.com/21985684/235062826-bf1391d0-374a-44c1-a7e3-0e9236eb6ae4.png) Replace all `torch.device('mps')` in the project. torch.device('mps') -> torch.device('cpu')

The MPS backend is supported on MacOS 12.3+.Current OS version can be queried using `sw_vers`

@moonskys all files. you need to search all files and replace. torch.device('mps') -> torch.device('cpu')

Caching for chat based models.

it will be fixed, detail: https://github.com/zilliztech/GPTCache/pull/538

Add the `GPTCacher` programming tool

https://github.com/zilliztech/GPTCache sorry, i wrongly write

Add the `GPTCacher` programming tool

@EwingYangs fix it

Have resolved：GPTcache :[Errno 63] File name too long: "similar_cache_[('_type', 'openai'), ('best_of', 2), ('frequency_penalty', 0), ('logit_bias', {}), ('max_tokens', 256), ('model_name', 'text-davinci-002'), ('n', 2), ('presence_penalty', 0), ('request_timeout', None), ('stop', None), ('temperature', 0.7), ('top_p', 1)]

great, thanks your attention, the llm parameter seems to have been modified by someone.