Commits

Branch selector

 main

User selector

Isotr0py

All time

Commit History

Commits on Mar 1, 2025

• [Doc] Consolidate `whisper` and `florence2` examples (#14050)

  Isotr0py

  authored
  fdcc405

  Copy full SHA for fdcc405

Commits on Feb 27, 2025

• [Bugfix] Fix qwen2.5-vl overflow issue (#13968)

  Show description for 7864875
  Isotr0py

  authored
  7864875

  Copy full SHA for 7864875

• [VLM] Support multimodal inputs for Florence-2 models (#13320)

  Isotr0py

  authored
  edf309e

  Copy full SHA for edf309e

Commits on Feb 25, 2025

• [Bugfix] Fix deepseek-vl2 inference with more than 2 images (#13818)

  Isotr0py

  authored
  6ff5186

  Copy full SHA for 6ff5186

Commits on Feb 23, 2025

• [LMM] Implement merged multimodal processor for whisper (#13278)

  Isotr0py

  authored
  ba5106e

  Copy full SHA for ba5106e

Commits on Feb 22, 2025

• [Misc] Fix yapf linting tools etc not running on pre-commit (#13695)

  Show description for ada7c78
  Isotr0py

  authored
  ada7c78

  Copy full SHA for ada7c78

Commits on Feb 21, 2025

• [Bugfix][CPU] Fix cpu all-reduce using native pytorch implementation (#13586)

  Isotr0py

  authored
  b2c3fc5

  Copy full SHA for b2c3fc5

Commits on Feb 19, 2025

• [Misc] Avoid calling unnecessary `hf_list_repo_files` for local model path (#13348)

  Show description for 550d97e
  Isotr0py

  authored
  550d97e

  Copy full SHA for 550d97e

Commits on Feb 18, 2025

• [Bugfix] Remove noisy error logging during local model loading (#13458)

  Isotr0py

  authored
  29fc577

  Copy full SHA for 29fc577

• [Bugfix] Fix failing transformers dynamic module resolving with spawn multiproc method (#13403)

  Show description for 8cf97f8
  Isotr0py

  authored
  8cf97f8

  Copy full SHA for 8cf97f8

• [Model] Enable quantization support for `transformers` backend (#12960)

  Isotr0py

  authored
  67ef8f6

  Copy full SHA for 67ef8f6

Commits on Feb 16, 2025

• [Bugfix][Platform][CPU] Fix cuda platform detection on CPU backend edge case (#13358)

  Show description for d67cc21
  Isotr0py

  authored
  d67cc21

  Copy full SHA for d67cc21

Commits on Feb 15, 2025

• [Bugfix] Fix qwen2.5-vl image processor (#13286)

  Isotr0py

  authored
  7fdaaf4

  Copy full SHA for 7fdaaf4

Commits on Feb 13, 2025

• [VLM] Remove input processor from clip and siglip (#13165)

  Isotr0py

  authored
  fa253f1

  Copy full SHA for fa253f1

• [VLM] Implement merged multimodal processor for Mllama (#11427)

  Isotr0py

  authored
  bc55d13

  Copy full SHA for bc55d13

Commits on Feb 8, 2025

• [Misc] Add qwen2.5-vl BNB support (#12944)

  Isotr0py

  authored
  4c8dd12

  Copy full SHA for 4c8dd12

Commits on Feb 6, 2025

• [Bugfix] Fix unsupported FA version check for Turing GPU (#12828)

  Isotr0py

  authored
  8108ac8

  Copy full SHA for 8108ac8

• [Kernel] Make rotary_embedding ops more flexible with input shape (#12777)

  Isotr0py

  authored
  85ac82d

  Copy full SHA for 85ac82d

Commits on Feb 5, 2025

• [VLM] Add MLA with pure RoPE support for deepseek-vl2 models (#12729)

  Isotr0py

  authored
  98fd089

  Copy full SHA for 98fd089

Commits on Feb 4, 2025

• [VLM] merged multimodal processor and V1 support for idefics3 (#12660)

  Show description for 815079d
  

  Isotr0py

  and

  DarkLight1337

  authored
  815079d

  Copy full SHA for 815079d

Commits on Jan 27, 2025

• [Bugfix] Fix gpt2 GGUF inference (#12467)

  Show description for ce69f7f
  Isotr0py

  authored
  ce69f7f

  Copy full SHA for ce69f7f

• [Bugfix] Fix missing seq_start_loc in xformers prefill metadata (#12464)

  Show description for 372bf08
  Isotr0py

  authored
  372bf08

  Copy full SHA for 372bf08

Commits on Jan 25, 2025

• [Misc] Add FA2 support to ViT MHA layer (#12355)

  Show description for f1fc051
  Isotr0py

  authored
  f1fc051

  Copy full SHA for f1fc051

Commits on Jan 23, 2025

• [Docs] Document Phi-4 support (#12362)

  Show description for 2cbeeda
  Isotr0py

  authored
  2cbeeda

  Copy full SHA for 2cbeeda

• [Bugfix] Fix broken internvl2 inference with v1 (#12360)

  Show description for 8c01b80
  Isotr0py

  authored
  8c01b80

  Copy full SHA for 8c01b80

• [Bugfix] Fix k_proj's bias for whisper self attention (#12342)

  Show description for c5b4b11
  Isotr0py

  authored
  c5b4b11

  Copy full SHA for c5b4b11

Commits on Jan 20, 2025

• [Model] Add Qwen2 PRM model support (#12202)

  Show description for 8360979
  Isotr0py

  authored
  8360979

  Copy full SHA for 8360979

Commits on Jan 19, 2025

• [Misc] Add BNB support to GLM4-V model (#12184)

  Show description for edaae19
  Isotr0py

  authored
  edaae19

  Copy full SHA for edaae19

Commits on Jan 18, 2025

• [Model] Port deepseek-vl2 processor, remove dependency (#12169)

  Show description for 02798ec
  Isotr0py

  authored
  02798ec

  Copy full SHA for 02798ec

Commits on Jan 17, 2025

• [Misc] Add deepseek_vl2 chat template (#12143)

  Show description for d75ab55
  Isotr0py

  authored
  d75ab55

  Copy full SHA for d75ab55

Commits on Jan 16, 2025

• [Model] Add support for deepseek-vl2-tiny model (#12068)

  Show description for 62b06ba
  Isotr0py

  authored
  62b06ba

  Copy full SHA for 62b06ba

• [Bugfix] Remove hardcoded `head_size=256` for Deepseek v2 and v3 (#12067)

  Show description for dd7c9ad
  Isotr0py

  authored
  dd7c9ad

  Copy full SHA for dd7c9ad

Commits on Jan 13, 2025

• [Model] Support GGUF models newly added in `transformers` 4.46.0 (#9685)

  Show description for d14e98d
  

  Isotr0py

  and

  DarkLight1337

  authored
  d14e98d

  Copy full SHA for d14e98d

Commits on Jan 12, 2025

• [Model] Initialize support for Deepseek-VL2 models (#11578)

  Show description for f967e51
  

  Isotr0py

  and

  DarkLight1337

  authored
  f967e51

  Copy full SHA for f967e51

Commits on Jan 10, 2025

• [Misc] Clean up debug code in Deepseek-V3 (#11930)

  Show description for 46fa98c
  Isotr0py

  authored
  46fa98c

  Copy full SHA for 46fa98c