huggingface/text-generation-inference

Pull Requests Commits

[gaudi] Move the _update_cos_sin_cache into get_cos_sin (#3254)

yuanwu2017 committed 232 days ago

Verified 25fdc5f0

[gaudi] Vlm rebase and issue fix in benchmark test (#3263)

sywangyi committed 232 days ago

Verified 613b8dd6

[gaudi] Perf optimization (#3256)

sywangyi committed 234 days ago

Verified 83947767

Bump neuron SDK version (#3260)

dacorvo committed 234 days ago

Verified 79183d16

Remove useless packages (#3253)

yuanwu2017 committed 242 days ago

Verified 1ff9d185

Prepare for 3.3.2 (#3249)

danieldk committed 245 days ago

Verified 249189d9

[gaudi] Fix the Llama-4-Maverick-17B-128E crash issue (#3246)

yuanwu2017 committed 247 days ago

Verified 6b6e30a6

[Gaudi] Fix the OOM issue of Llama-4-Scout-17B-16E-Instruct (#3245)

yuanwu2017 committed 247 days ago

Verified 70217ac3

fp8 compressed tensors w8a8 support for Gaudi backend (#3242)

sywangyi committed 248 days ago

Verified f1404400

Add Qwen3 for Gaudi backend (#3229)

yuanwu2017 committed 253 days ago

Verified 1883a62a

Nix: switch to hf-nix (#3240)

danieldk committed 253 days ago

Verified f58d7cf5

Upgrade to new vllm extension ops for Gaudi backend (fix issue in exponential bucketing) (#3239)

sywangyi committed 254 days ago

Verified f08b44ad

Prepare for 3.3.1 (#3238)

danieldk committed 254 days ago

Verified 674c514d

Move input_ids to hpu and remove disposal of adapter_meta (#3237)

sywangyi committed 254 days ago

Verified 9e7e5469

Switch to punica-sgmv kernel from the Hub (#3236)

danieldk committed 254 days ago

Verified e3252879

Fix the crash in default ATTENTION path for Gaudi backend (#3235)

sywangyi committed 256 days ago

Verified 43b1b07f

Refine warmup and upgrade to synapse AI 1.21.0 (#3234)

sywangyi committed 256 days ago

Verified 000e313a

Deepseek R1 for Gaudi backend (#3211)

sywangyi committed 256 days ago

Verified d658b5de

fix: count gpu uuids if NVIDIA_VISIBLE_DEVICES env set to all (#3230)

drbh committed 259 days ago

Verified 58934c8b

Enable Llama4 for Gaudi backend (#3223)

yuanwu2017 committed 261 days ago

Verified 18cbecfb

Update to Torch 2.7.0 (#3221)

danieldk committed 261 days ago

Verified 7e531f41

Adjust the `round_up_seq` logic in Gaudi backend (#3224)

kaixuanliu committed 264 days ago

Verified 535ce238

Change HPU warmup logic: seq length should be with exponential growth (#3217)

kaixuanliu committed 265 days ago

Verified c94f415a

Prepare for 3.3.0 (#3220)

danieldk committed 266 days ago

Verified 56c81894

Chunked Prefill VLM (#3188)

mht-sharma committed 269 days ago

Verified 329f612e

forward and tokenize chooser use the same shape (#3196)

sywangyi committed 270 days ago

Verified 533eee50

IPEX support FP8 kvcache/softcap/slidingwindow (#3144)

sywangyi committed 270 days ago

Verified 51a0b9d1

Fix `HF_HUB_OFFLINE=1` for Gaudi backend (#3193)

regisss committed 270 days ago

Verified f208ba6a

Update client SDK snippets (#3207)

julien-c committed 274 days ago

Verified 7253be34

fix: bump snaps for mllama (#3202)

drbh committed 274 days ago

Verified d303c1e3

Newer Older