xenova/transformers.js

Pull Requests Commits

fix granite speech processor num features (#1685)

xenova committed 7 days ago

Verified 5ed4c5c0

Add support for Zaya models (#1682)

xenova committed 16 days ago

Verified d319c3d0

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

xenova committed 17 days ago

Verified 3dc58cbc

Add support for DeepSeek-V4 (#1679)

xenova committed 17 days ago

Verified 857b29a5

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

xenova committed 17 days ago

Verified 35636dff

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

anishesg committed 28 days ago

Verified f7487c73

[version] Update to 4.2.0

xenova committed 33 days ago

Verified 6688e09c

Add support for OpenAI privacy filter model (#1658)

xenova committed 33 days ago

Verified ea88f897

Use inputMetadata API for simplified internals (#1657)

xenova committed 33 days ago

Verified b93766d7

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

nico-martin committed 33 days ago

Verified 6eaf11a0

[version] Update to 4.1.0

xenova committed 40 days ago

Verified ca10e4d1

Update default generation parameters (#1649)

xenova committed 40 days ago

Verified fbab88f6

Re-enable SmolVLM (#1648)

pdufour committed 40 days ago

Verified 39bec702

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

rtrompier committed 40 days ago

Verified 78d4ec3d

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

xenova committed 40 days ago

Verified fdd496e9

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

xenova committed 41 days ago

Verified ed84c15f

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

xenova committed 42 days ago

Verified 8d615c5a

🔒 Pin GitHub Actions to commit SHAs (#1626)

paulinebm committed 48 days ago

Verified 60379e22

[version] Update to 4.0.1

xenova committed 53 days ago

Verified 64da6359

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

xenova committed 53 days ago

Verified 3cee1ac2

[version] Update to 4.0.0 (#1620)

xenova committed 56 days ago

Verified 364ebd41

[version] Update to 4.0.0-next.11

xenova committed 56 days ago

Verified 3458fa5e

Upgrade onnxruntime-web version (#1618)

xenova committed 56 days ago

Verified 28f92674

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

xenova committed 56 days ago

Verified ab22b27e

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

xenova committed 56 days ago

Verified d8340931

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

xenova committed 56 days ago

Verified a0d86d50

Overdue Whisper fixes (#1594)

xenova committed 57 days ago

Verified 43b26629

[version] Update to 4.0.0-next.10

xenova committed 59 days ago

Verified d6280f73

Add support for Cohere ASR (#1610)

xenova committed 59 days ago

Verified b368d75e

Fix typo in WebGPU documentation (#1607)

mattsoulanille committed 60 days ago

Verified e94d5592

Older