Kezdőlap Felfedezés Súgó
Regisztráció Bejelentkezés
david
/
aphrodite-engine
tükrözi: https://github.com/PygmalionAI/aphrodite-engine
1
0
Másolás 0
Fájlok Problémák 0 Wiki
Branch: lm_head_lora
Branch-ok Tag-ek
aphrodite-engin...
/
aphrodite
/
modeling
/
models
AlpinDale fc5ef786b0 Merge branch 'main' into lm_head_lora 1 hete
..
__init__.py a5bfc2bc3d VLM: add support for LLaVA-Onevision model (#1100) 1 hete
arctic.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
baichuan.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
bart.py a985143768 core: add cuda graph support for encoder-decoder models (#1051) 1 hónapja
blip.py 7ce3174039 VLM: refactor blip models to support composite weight loading (#1098) 1 hete
blip2.py 7ce3174039 VLM: refactor blip models to support composite weight loading (#1098) 1 hete
bloom.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
chameleon.py 651678d2df VLM: use `SequenceData.from_token_counts` to create dummy data (#1093) 1 hete
chatglm.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
clip.py a5bfc2bc3d VLM: add support for LLaVA-Onevision model (#1100) 1 hete
commandr.py 135dfd648b fix: LoRA support for Cohere and Jamba models (#1004) 1 hónapja
dbrx.py b65449b5ad moe: refactor DBRX experts to support FusedMoE (#1095) 1 hete
decilm.py 9022c6d869 remove progress_bar imports 3 hónapja
deepseek.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
deepseek_v2.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
eagle.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
exaone.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
falcon.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
fuyu.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
gemma.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
gemma2.py b33cf04386 quants: add bitsandbytes support for gemma2 model (#1026) 1 hónapja
gpt2.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
gpt_bigcode.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
gpt_j.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
gpt_neox.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
granite.py 11f49b5341 fix: granite logit scale in logit computation (#1054) 1 hónapja
idefics2_vision_model.py f1d0b77c92 [0.6.0] Release Candidate (#481) 5 hónapja
interfaces.py 80e7f3b0bd feat: support finetuned lm_head and embed_tokens in LoRA adapters 1 hónapja
intern_vit.py f56d6b396a vlm: fallback to SDPA for ViT models on CPU backend (#982) 1 hónapja
internlm2.py 7632f91429 fix: InternLM2 model with Tensor Parallel (#980) 1 hónapja
internvl.py 91d03c04d2 VLM: refactor composite weight loading logic (#1097) 1 hete
jais.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
jamba.py 135dfd648b fix: LoRA support for Cohere and Jamba models (#1004) 1 hónapja
llama.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
llama_embedding.py 9022c6d869 remove progress_bar imports 3 hónapja
llava.py 91d03c04d2 VLM: refactor composite weight loading logic (#1097) 1 hete
llava_next.py 7ce3174039 VLM: refactor blip models to support composite weight loading (#1098) 1 hete
llava_next_video.py 91d03c04d2 VLM: refactor composite weight loading logic (#1097) 1 hete
llava_onevision.py a5bfc2bc3d VLM: add support for LLaVA-Onevision model (#1100) 1 hete
mamba.py 3bb0f07461 chore: rename `task_handler` to `worker` (#985) 1 hónapja
mamba_cache.py a113309876 kernel: add meta functions for ops to prevent graph breaks (#1019) 1 hónapja
medusa.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
minicpm.py ce7b602f03 model: add support for MiniCPM-3 (#1044) 1 hónapja
minicpm3.py ce7b602f03 model: add support for MiniCPM-3 (#1044) 1 hónapja
minicpmv.py 651678d2df VLM: use `SequenceData.from_token_counts` to create dummy data (#1093) 1 hete
mixtral.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
mixtral_quant.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
mlp_speculator.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
molmo.py acc0c727c8 vlm: add support for molmo vision model (#1069) 1 hónapja
mpt.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
na_vit.py 9f3e7c86e2 feat: add fused Marlin MoE kernel (#934) 1 hónapja
nemotron.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
olmo.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
olmoe.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
opt.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
orion.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
paligemma.py 91d03c04d2 VLM: refactor composite weight loading logic (#1097) 1 hete
persimmon.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
phi.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
phi3.py ec17b6c4d0 fix: Phi3.5 Mini and MoE LoRA inference (#1070) 1 hónapja
phi3_small.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
phi3v.py 4d14bd1fe5 vlm: add multi-input support for LLaVA and InternVL models (#1002) 1 hónapja
phimoe.py ec17b6c4d0 fix: Phi3.5 Mini and MoE LoRA inference (#1070) 1 hónapja
pixtral.py 651678d2df VLM: use `SequenceData.from_token_counts` to create dummy data (#1093) 1 hete
qwen.py 651678d2df VLM: use `SequenceData.from_token_counts` to create dummy data (#1093) 1 hete
qwen2.py c90abcc603 VLM: add pipeline parallelism support for Qwen2-VL (#1103) 1 hete
qwen2_moe.py 5224389dae chore: skip loading extra bias for qwen2 moe GPTQ (#1011) 1 hónapja
qwen2_vl.py c90abcc603 VLM: add pipeline parallelism support for Qwen2-VL (#1103) 1 hete
siglip.py a5bfc2bc3d VLM: add support for LLaVA-Onevision model (#1100) 1 hete
solar.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
stablelm.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
starcoder2.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja
ultravox.py 91d03c04d2 VLM: refactor composite weight loading logic (#1097) 1 hete
utils.py 91d03c04d2 VLM: refactor composite weight loading logic (#1097) 1 hete
xverse.py 0dfa6b60ec core: support logprobs with multi-step scheduling (#963) 1 hónapja