5.37 GB

Ctrl+K

1 contributor

History: 38 commits

erik-at-boson

Update weights: MERGE_PROVENANCE.json

2ffd1aa verified 4 months ago

open_asr_leaderboard
Add leaderboard reproduction helpers 6 months ago
.gitattributes

1.57 kB
Upload folder using huggingface_hub 7 months ago
MERGE_PROVENANCE.json

685 Bytes
Update weights: MERGE_PROVENANCE.json 4 months ago
README.md

6.02 kB
Update weights: README.md 4 months ago
added_tokens.json

1.19 kB
Upload folder using huggingface_hub 7 months ago
attention.py

8.71 kB
Fix: use importlib for deepspeed to avoid static import check 7 months ago
common.py

1.09 kB
fix: dual-mode imports + complete ChatMLDatasetSample + transcribe output parsing 6 months ago
config.json

5.91 kB
fix: add auto_map + make xcodec imports lazy for trust_remote_code support 6 months ago
configuration_higgs_audio.py

9.88 kB
Fix: handle higgs_audio_encoder config without global registry 7 months ago
cuda_graph_runner.py

5.14 kB
Add model code for trust_remote_code support 7 months ago
custom_modules.py

6.19 kB
Add model code for trust_remote_code support 7 months ago
generation_config.json

147 Bytes
Upload folder using huggingface_hub 7 months ago
higgs_audio_collator.py

36.6 kB
fix: dual-mode imports + complete ChatMLDatasetSample + transcribe output parsing 6 months ago
merges.txt

1.67 MB
Upload folder using huggingface_hub 7 months ago
model-00001-of-00002.safetensors

4.73 GB
xet

Update weights: model-00001-of-00002.safetensors 4 months ago
model-00002-of-00002.safetensors

622 MB
xet

Upload folder using huggingface_hub 7 months ago
model.safetensors.index.json

65.7 kB
Upload folder using huggingface_hub 7 months ago
modeling_higgs_audio.py

110 kB
feat: batched generation for ASR (remove bs=1 assertion when no audio_out tokens) 6 months ago
modeling_higgs_audio_xcodec.py

20.2 kB
fix: dual-mode imports + complete ChatMLDatasetSample + transcribe output parsing 6 months ago
ngram_loop_fix.py

2.07 kB
Update v2: ngram_loop_fix.py 4 months ago
special_tokens_map.json

613 Bytes
Upload folder using huggingface_hub 7 months ago
tokenizer.json

11.4 MB
xet

Upload folder using huggingface_hub 7 months ago
tokenizer_config.json

12.7 kB
Upload folder using huggingface_hub 7 months ago
transcribe.py

16.1 kB
transcribe.py: inline n-gram loop fix (harness loader fetches a fixed sibling list) 4 months ago
utils.py

36.5 kB
Fix: use importlib for deepspeed to avoid static import check 7 months ago
vocab.json

2.78 MB
Upload folder using huggingface_hub 7 months ago