huggingface/text-generation-inference

Pull Requests Commits

danieldk committed 211 days ago

767a6520

Prepare for 3.3.1 (#3238)

danieldk committed 211 days ago

Verified 674c514d

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

sywangyi committed 211 days ago

Verified 9e7e5469

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

danieldk committed 211 days ago

Verified e3252879

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

sywangyi committed 212 days ago

Verified 43b1b07f

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

sywangyi committed 213 days ago

Verified 000e313a

Deepseek R1 for Gaudi backend (#3211)

sywangyi committed 213 days ago

Verified d658b5de

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

drbh committed 216 days ago

Verified 58934c8b

Enable Llama4 for Gaudi backend (#3223)

yuanwu2017 committed 217 days ago

Verified 18cbecfb

Update to Torch 2.7.0 (#3221)

danieldk committed 218 days ago

Verified 7e531f41

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

kaixuanliu committed 221 days ago

Verified 535ce238

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

kaixuanliu committed 222 days ago

Verified c94f415a

Prepare for 3.3.0 (#3220)

danieldk committed 223 days ago

Verified 56c81894

Chunked Prefill VLM (#3188)

mht-sharma committed 226 days ago

Verified 329f612e

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

sywangyi committed 227 days ago

Verified 533eee50

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

sywangyi committed 227 days ago

Verified 51a0b9d1

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

regisss committed 227 days ago

Verified f208ba6a

Update client SDK snippets (#3207)

julien-c committed 231 days ago

Verified 7253be34

fix: bump snaps for mllama (#3202)

drbh committed 231 days ago

Verified d303c1e3

Pr 2982 ci branch (#3046)

drbh committed 231 days ago

Verified 12ea8d74

doc typo (#3206)

julien-c committed 231 days ago

Verified 6afe4307

Skip `{% generation %}` and `{% endgeneration %}` template handling (#3204)

alvarobartt committed 232 days ago

Verified 40dfce64

Fixing the router + template for Qwen3. (#3200)

Narsil committed 233 days ago

Verified e7329fec

Put more wiggle room. (#3189)

Narsil committed 238 days ago

Verified 39cfe232

Warmup gaudi backend (#3172)

sywangyi committed 239 days ago

Verified 37580294

Add option to configure prometheus port (#3187)

mht-sharma committed 239 days ago

Verified 02715dc5

Fixing CI (#3184)

Narsil committed 245 days ago

Verified 8f881979

Bump `sccache` to 0.10.0 (#3179)

alvarobartt committed 245 days ago

Verified 95ccba37

Get opentelemetry trace id from request headers instead of creating a new trace (#2648)

kozistr committed 245 days ago

Verified b400c275

Support flashinfer for Gemma3 prefill (#3167)

danieldk committed 245 days ago

Verified 84ab88d8

Older