Daniël de Kok c6071749db Fix mask passed to flashinfer (#3324) ... Custom masks are padded to the shape `[batch_size, max_len, max_len]`. However, flashinfer expects an unpadded mask of the shape `[sum(q_len[i] * k_len[i] for i in range(batch_size)]`. This change unpads the custom mask (currently only used by Gemma 3) to this shape (assuming q_len == k_len, since we only use the custom mask during prefill).		2025-09-08 13:47:03 -04:00
..
adapters	xpu lora support (#3232)	2025-07-02 17:54:25 +02:00
layers	Fix mask passed to flashinfer (#3324)	2025-09-08 13:47:03 -04:00
models	chore: prepare version 3.3.5 (#3314)	2025-09-02 15:35:42 +02:00
pb	chore: add pre-commit (#1569)	2024-02-16 11:58:58 +01:00
utils	chore: prepare version 3.3.5 (#3314)	2025-09-02 15:35:42 +02:00
__init__.py	feat(clients): Python client (#103)	2023-03-07 18:52:22 +01:00
cache.py	fix(server): decrease memory fragmentation (#557)	2023-07-06 14:28:33 +02:00
cli.py	Fixing TRTLLM dockerfile. (#2922)	2025-01-20 11:13:46 +01:00
interceptor.py	feat: prefill chunking (#2600)	2024-10-16 12:49:33 +02:00
server.py	Chunked Prefill VLM (#3188)	2025-05-06 18:01:59 +02:00
tracing.py	Add OTLP Service Name Environment Variable (#2076)	2024-06-25 09:33:01 +02:00