Sanchit Gandhi

Un-pin JAX from <= 0.3.6

2

### Feature request JAX was pinned to

Whisper Prompting

8

### Feature request Add prompting for the Whisper model to control the style/formatting of the generated text. ### Motivation During training, Whisper can be fed a "previous context window" to...

Split dataset by node: index error when sharding iterable dataset

3

### Describe the bug Context: we're splitting an iterable dataset by node and then passing it to a torch data loader with multiple workers When we iterate over it for...

Streaming dataset looses `.feature` method after `.add_column`

2

### Describe the bug After appending a new column to a streaming dataset using `.add_column`, we can no longer access the list of dataset features using the `.feature` method. ###...

Audio Streaming: large latency before first chunk is played

2

We typically stream audio outputs when latency is a major consideration. E.g. if we're generating 10-seconds of audio and want the perceived latency to be as low as possible, we...

### Describe the bug Using the microphone input on Spaces is temperamental, often forwarding empty inputs to our function. Response from @abidlabs: > So my hunch is that the audio...

If we copy the URL for generate parameters that are formatted as: `parameter(type hint)`, the type hint is copied into the URL link. For example, on the [Generation doc page](https://huggingface.co/docs/transformers/main_classes/text_generation),...

[MusicGen] SDPA gives nans/infs during sampling

3

### System Info - `transformers` version: 4.40.0.dev0 - Platform: Linux-5.4.0-166-generic-x86_64-with-glibc2.29 - Python version: 3.8.10 - Huggingface_hub version: 0.22.1 - Safetensors version: 0.4.2 - Accelerate version: 0.27.2 - Accelerate config: -...