Quentin Fuxa a540a5fd10 fix simul-kv audio trim bug, add 1.7B v2 alignment heads

2026-03-15 20:45:00 +01:00

..

alignment_heads_qwen3_asr_0.6B.json

qwen3 simul+kv: optimized streaming with kv cache reuse

2026-03-15 18:30:00 +01:00

alignment_heads_qwen3_asr_1.7B_v2.json

fix simul-kv audio trim bug, add 1.7B v2 alignment heads

2026-03-15 20:45:00 +01:00

alignment_heads_qwen3_asr_1.7B.json

qwen3 alignment heads

2026-03-14 00:12:50 +01:00

alignment_heads_qwen3_asr_1.7B.png

qwen3 alignment heads

2026-03-14 00:12:50 +01:00

alignment_heads.png

Add a script to detect alignement heads, usefull for distilled whisper

2025-11-09 18:12:09 +01:00

convert_hf_whisper.py

correct silences handling when language not auto

2025-11-20 11:20:00 +01:00

create_long_samples.py

update benchmark results and procedure

2026-03-15 11:16:15 +01:00

detect_alignment_heads_qwen3.py

qwen3 alignment heads

2026-03-14 00:12:50 +01:00

determine_alignment_heads.py

Lint scripts and tests

2026-01-04 14:15:00 +01:00

generate_architecture.py

update benchmark results and procedure

2026-03-15 11:16:15 +01:00

python_support_matrix.py

feat(tests): python matrix support test

2026-02-25 21:35:41 -03:00

run_scatter_benchmark.py

update benchmark with qwen3 which reuses kv cache

2026-03-15 22:32:01 +01:00

sync_extension.py

Lint scripts and tests

2026-01-04 14:15:00 +01:00