sujithjoseph

[BUG] ValueError: max() arg is an empty sequence using bf16 zero stage3

14

``` │ /opt/conda/lib/python3.7/site-packages/deepspeed/runtime/zero/stage3.py:307 │ │ in │ │ │ │ 304 │ │ │ max([ │ │ 305 │ │ │ │ max(tensor.numel(), │ │ 306 │ │ │ │...

How to save / form the config.json after fine-tuning - Flan T5 11b

46

After fine-tuning a flan t5 11b model on custom data, I was saving the checkpoint via accelerate like this ``` accelerator.wait_for_everyone() accelerator.save( get_peft_model_state_dict(model, state_dict=accelerator.get_state_dict(model)), checkpoint_name ) accelerator.wait_for_everyone() ``` It didnt...

### Feature request https://platform.openai.com/docs/api-reference/embeddings/create?lang=python supports a user parameter, where we can pass user details to the openai api. https://github.com/hwchase17/langchain/blob/master/langchain/embeddings/openai.py#L66 can take a user parameter (optional) , which needs to be...

### Feature request Can this API support Assisted Generation feature from https://huggingface.co/blog/assisted-generation? ### Motivation Better Latency. ### Your contribution Yes..

Support for Falcon-7B / 40B models

5

It would be great, if you can add support for Falcon models as well! Does it support onnx models today?

LongT5 support

1

### Model description Will the current T5 loader support text gen using LongT5 as well (https://github.com/huggingface/text-generation-inference/blob/main/server/text_generation_server/models/t5.py) . ### Open source status - [X] The model implementation is available - [X]...