dengxuan
968d800110
fix(embed): to_dtype F32 before to_vec1以适应 F16 模型
ci / cargo fmt (push) Failing after 1m45s
ci / cargo clippy (push) Failing after 4m20s
ci / build and test (push) Successful in 4m33s
2026-07-25 14:12:07 +08:00
dengxuan
2df0929440
feat(embed): MiniCPM4 embed/embed_batch + MiniCPM4GenerateModel embed_text/embed_text_batch
...
ci / cargo fmt (push) Failing after 2m43s
ci / cargo clippy (push) Failing after 5m9s
ci / build and test (push) Successful in 8m32s
- MiniCPMModel::embed: run forward through layers + norm, mean pool, L2 normalize
- MiniCPMModel::embed_batch: batch version
- MiniCPM4GenerateModel::embed_text: tokenize + embed + return Vec<f32>
- MiniCPM4GenerateModel::embed_text_batch: batch version
2026-07-25 13:03:25 +08:00
jhqxxx
257057de07
generate code refactoring progress 1/3
2026-05-28 23:21:15 +08:00
jhqxxx
35678e3ee4
add MiniCPM5-1B
2026-05-27 00:50:23 +08:00
jhqxxx
279480e3d7
generate add penalty repeat
2026-04-03 14:15:10 +08:00
jhqxxx
bd8eee6520
refactor generate code
2026-04-02 22:22:52 +08:00
jhqxxx
b254d21efc
refactor deepseek_ocr/fun_asr_nano generate code
2026-04-01 19:11:10 +08:00
jhqxxx
3e510a4172
dependencies delete aha_openai_dive,chrono
2026-03-31 12:30:12 +08:00
jhqxxx
363833d8fb
delete some code
2026-03-30 20:36:57 +08:00
jhqxxx
ccbe799b90
update model_name
2026-03-30 20:33:48 +08:00
jhqxxx
037207eccd
update generate param init
2026-03-06 19:08:58 +08:00
jhqxxx
548eb8185e
usage add prompt_tokens
2026-03-06 18:38:16 +08:00
YageGeng
42c95da0e3
refactor: remove unstable Option<&T> branch by returning owned Tensor
2025-12-11 10:50:35 +08:00
jhqxxx
1a0df13bc7
model test add tps
2025-12-10 16:44:20 +08:00
jhqxxx
697484cf23
add hunyuan_ocr
2025-12-03 17:21:01 +08:00
jhqxxx
d15cd45315
add deploy model
2025-11-05 14:46:03 +08:00
jhqxxx
1dcccf5aeb
update dependences
2025-10-27 10:55:40 +08:00
jhqxxx
7b190e7d54
add Qwen3VL model
2025-10-26 21:39:23 +08:00
Yijun Zhao
9b9a8f2c73
unify cargo version and add some ci rules
2025-10-15 21:03:49 +08:00
jhqxxx
39536df06b
update voxcpm
2025-10-10 20:36:52 +08:00
jhqxxx
f33eaaee0d
add voxcpm with some bug
2025-10-03 22:25:58 +08:00
jhqxxx
ae36194a4b
update init
2025-09-25 12:44:17 +08:00
jhqxxx
4fdab3e7cd
add minicpm with a bug
2025-09-25 12:09:25 +08:00