Firworks commited on
Commit
130497c
·
verified ·
1 Parent(s): 935276f

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +0 -13
README.md CHANGED
@@ -42,25 +42,12 @@ It does not generate audio; it produces text based on audio input.
42
  # 📦 What This Quantized Version Enables
43
 
44
  This NVFP4 quantized version reduces memory requirements significantly:
45
-
46
  Size: ~22 GB (down from ~67 GB)
47
-
48
  Should fit comfortably on a single RTX 5090
49
 
50
- Fully compatible with vLLM (including streaming text output)
51
-
52
  Preserves most reasoning performance from the BF16 release
53
-
54
  Because of this, anyone with a high-end consumer GPU can experiment with advanced audio reasoning locally.
55
 
56
- 🖥 Supported Audio Behavior
57
-
58
- The model supports:
59
-
60
- ✔ Streaming text output through vLLM
61
- ✔ Reading uploaded audio files (WAV/MP3/etc) via ffmpeg
62
- ✘ It does not synthesize audio
63
- ✘ It does not require pre-burned waveforms — any user-provided audio file works
64
 
65
  Check the original model card for more information about this model.
66
 
 
42
  # 📦 What This Quantized Version Enables
43
 
44
  This NVFP4 quantized version reduces memory requirements significantly:
 
45
  Size: ~22 GB (down from ~67 GB)
 
46
  Should fit comfortably on a single RTX 5090
47
 
 
 
48
  Preserves most reasoning performance from the BF16 release
 
49
  Because of this, anyone with a high-end consumer GPU can experiment with advanced audio reasoning locally.
50
 
 
 
 
 
 
 
 
 
51
 
52
  Check the original model card for more information about this model.
53