xenova/transformers.js

Pull Requests Commits

Disable asyncify for Safari below 26 when WebGPU is not available

xenova committed 146 days ago

087ae81c

Enable WebGPU for Safari 26 and above

xenova committed 147 days ago

d9f6a692

Add support for HRM-Text (#1695)

xenova committed 148 days ago

Verified bc9cf740

Fix missing global exports (#1696)

xenova committed 149 days ago

Verified 7259f84e

fix granite speech processor num features (#1685)

xenova committed 166 days ago

Verified 5ed4c5c0

Add support for Zaya models (#1682)

xenova committed 174 days ago

Verified d319c3d0

Ensure `num_logits_to_keep` is always set to 1 for generate calls (#1681)

xenova committed 175 days ago

Verified 3dc58cbc

Add support for DeepSeek-V4 (#1679)

xenova committed 175 days ago

Verified 857b29a5

Fix progress callbacks for pipelines that load files multiple times (#1680)

xenova committed 175 days ago

Verified 35636dff

Fix duplicate model file downloads when progress_callback is active (#1663) (#1664)

anishesg committed 186 days ago

Verified f7487c73

[version] Update to 4.2.0

xenova committed 191 days ago

Verified 6688e09c

Add support for OpenAI privacy filter model (#1658)

xenova committed 191 days ago

Verified ea88f897

Use inputMetadata API for simplified internals (#1657)

xenova committed 191 days ago

Verified b93766d7

Added `tools` to `TextGenerationPipeline` (#1655)

nico-martin committed 192 days ago

Verified 6eaf11a0

[version] Update to 4.1.0

xenova committed 198 days ago

Verified ca10e4d1

Update default generation parameters (#1649)

xenova committed 198 days ago

Verified fbab88f6

Re-enable SmolVLM (#1648)

pdufour committed 198 days ago

Verified 39bec702

chore: bump doc-builder SHA for PR upload workflow (#1644)

rtrompier committed 198 days ago

Verified 78d4ec3d

Add support for q1, q1f16, q2, q2f16 data types (#1647)

xenova committed 198 days ago

Verified fdd496e9

Improve tokenizer types based on input function parameters (#1641)

xenova committed 199 days ago

Verified ed84c15f

Cached generation improvements (+ past_key_values via pipeline function) (#1638)

xenova committed 200 days ago

Verified 8d615c5a

🔒 Pin GitHub Actions to commit SHAs (#1626)

paulinebm committed 206 days ago

Verified 60379e22

[version] Update to 4.0.1

xenova committed 211 days ago

Verified 64da6359

casually dropping the most capable open weights on the planet (#1627)

xenova committed 211 days ago

Verified 3cee1ac2

[version] Update to 4.0.0 (#1620)

xenova committed 214 days ago

Verified 364ebd41

[version] Update to 4.0.0-next.11

xenova committed 215 days ago

Verified 3458fa5e

Upgrade onnxruntime-web version (#1618)

xenova committed 215 days ago

Verified 28f92674

Fix custom cache in Node.js when cache returns non-paths (e.g., Response) (#1617)

xenova committed 215 days ago

Verified ab22b27e

Emit `progress_total` events from `PreTrainedModel.from_pretrained()` (#1615)

xenova committed 215 days ago

Verified d8340931

Fix ModelRegistry calls when loading multimodal models with text-generation pipeline (#1614)

xenova committed 215 days ago

Verified a0d86d50

Older