 yuiseki
|
5d5c066de8
mtmd : fix Pixtral OOM with large images by capping image_size to 1024 (#14326)
|
hace 7 meses |
 Xuan-Son Nguyen
|
413977de32
mtmd : refactor llava-uhd preprocessing logic (#14247)
|
hace 7 meses |
|
Xuan-Son Nguyen
|
10961339b2
mtmd : move helpers to dedicated library (⚠️ breaking change) (#13866)
|
hace 7 meses |
|
Xuan-Son Nguyen
|
bc583e3c63
mtmd : support Qwen 2.5 Omni (input audio+vision, no audio output) (#13784)
|
hace 7 meses |
|
Xuan-Son Nguyen
|
40aaa8a403
mtmd : add support for Qwen2-Audio and SeaLLM-Audio (#13760)
|
hace 7 meses |
|
Xuan-Son Nguyen
|
797990c4bc
mtmd : add ultravox audio input (#13623)
|
hace 8 meses |
|
Xuan-Son Nguyen
|
92ecdcc06a
mtmd : add vision support for llama 4 (#13282)
|
hace 8 meses |
|
Xuan-Son Nguyen
|
71bdbdb587
clip : clip.h become private API (⚠️ breaking change) (#13510)
|
hace 8 meses |
|
Xuan-Son Nguyen
|
b4726345ac
mtmd : remove libllava, remove clip-quantize-cli (⚠️ breaking change) (#13460)
|
hace 8 meses |
|
Xuan-Son Nguyen
|
de4c07f937
clip : cap max image size 1024 for qwen vl model (#13478)
|
hace 8 meses |
 City
|
c104023994
mtmd : Use RMS norm for InternVL 3 38B and 78B mmproj (#13459)
|
hace 8 meses |
 David Huang
|
7f323a589f
Add `--no-op-offload` to improve `-ot` pp perf in MoE models like llama4 400B (#13386)
|
hace 8 meses |
|
City
|
3eac209319
mtmd : support InternVL 3 38B and 78B mmproj (#13443)
|
hace 8 meses |
|
Xuan-Son Nguyen
|
15e6125a39
mtmd : add hard limit on image resolution for qwen2vl / qwen2.5vl (#13434)
|
hace 8 meses |
|
Xuan-Son Nguyen
|
053367d149
mtmd : support InternVL 2.5 and 3 (#13422)
|
hace 8 meses |
 Diego Devesa
|
27ebfcacba
llama : do not crash if there is no CPU backend (#13395)
|
hace 8 meses |
 welix
|
0ccc121354
mtmd : fix the calculation of n_tokens for smolvlm (#13381)
|
hace 8 meses |
|
Xuan-Son Nguyen
|
32916a4907
clip : refactor graph builder (#13321)
|
hace 8 meses |
|
Xuan-Son Nguyen
|
9b61acf060
mtmd : rename llava directory to mtmd (#13311)
|
hace 8 meses |
