Alcpz commited on
Commit
fb636ba
Β·
1 Parent(s): 7576abe

Reduced binary size, updated model card

Browse files
README.md CHANGED
@@ -60,11 +60,6 @@ runners
60
  β”‚ └── lfm2-audio-ubuntu-arm64.zip
61
  └── ubuntu-x64
62
  └── lfm2-audio-ubuntu-x64.zip
63
- β”‚Β Β  └── llama-mtmd-cli
64
- └── ubuntu-x64
65
- └── bin
66
- β”œβ”€β”€ llama-lfm2-audio
67
- └── llama-mtmd-cli
68
  ```
69
 
70
  Each package contains `llama-lfm2-audio` and `llama-mtmd-cli` binaries.
@@ -83,14 +78,14 @@ The mode is defined by system prompt. There are limitations on system prompt and
83
  ASR requires `-sys "Perform ASR."` and `--audio audio.wav` for input. It will print text to console
84
 
85
  ```console
86
- bin/llama-lfm2-audio -m $CKPT/LFM2-Audio-1.5B-Q8_0.gguf --mmproj $CKPT/mmproj-audioencoder-LFM2-Audio-1.5B-Q8_0.gguf -mv $CKPT/audiodecoder-LFM2-Audio-1.5B-Q8_0.gguf -sys "Perform ASR." --audio $INPUT_WAV
87
  ```
88
 
89
  ### TTS
90
 
91
  TTS requires `-sys "Perform TTS."`, `-p "What is this obsession people have with books?"` for input, and `--output output.wav` for output. It will save audio to `output.wav`.
92
  ```console
93
- bin/llama-lfm2-audio -m $CKPT/LFM2-Audio-1.5B-Q8_0.gguf --mmproj $CKPT/mmproj-audioencoder-LFM2-Audio-1.5B-Q8_0.gguf -mv $CKPT/audiodecoder-LFM2-Audio-1.5B-Q8_0.gguf -sys "Perform TTS." -p "What is this obsession people have with books?" --output $OUTPUT_WAV
94
  ```
95
 
96
  ### Interleaved
@@ -98,7 +93,7 @@ bin/llama-lfm2-audio -m $CKPT/LFM2-Audio-1.5B-Q8_0.gguf --mmproj $CKPT/mmproj-au
98
  Interleaved produces both, text and audio as output, and can consume text or audio as input.
99
 
100
  ```console
101
- bin/llama-lfm2-audio -m $CKPT/LFM2-Audio-1.5B-Q8_0.gguf --mmproj $CKPT/mmproj-audioencoder-LFM2-Audio-1.5B-Q8_0.gguf -mv $CKPT/audiodecoder-LFM2-Audio-1.5B-Q8_0.gguf -sys "Respond with interleaved text and audio." --audio $INPUT_WAV --output $OUTPUT_WAV
102
  ```
103
 
104
  ## Run ASR using `llama-mtmd-cli`
@@ -106,7 +101,7 @@ bin/llama-lfm2-audio -m $CKPT/LFM2-Audio-1.5B-Q8_0.gguf --mmproj $CKPT/mmproj-au
106
  Build `llama-mtmd-cli` following the standard build procedure.
107
 
108
  ```console
109
- bin/llama-mtmd-cli -m $CKPT/LFM2-Audio-1.5B-Q8_0.gguf --mmproj $CKPT/mmproj-audioencoder-LFM2-Audio-1.5B-Q8_0.gguf -p "<__media__>" -sys "Perform ASR." --audio $INPUT_WAV
110
  ```
111
 
112
  ### Debug
 
60
  β”‚ └── lfm2-audio-ubuntu-arm64.zip
61
  └── ubuntu-x64
62
  └── lfm2-audio-ubuntu-x64.zip
 
 
 
 
 
63
  ```
64
 
65
  Each package contains `llama-lfm2-audio` and `llama-mtmd-cli` binaries.
 
78
  ASR requires `-sys "Perform ASR."` and `--audio audio.wav` for input. It will print text to console
79
 
80
  ```console
81
+ lfm2-audio-<platform>/llama-lfm2-audio -m $CKPT/LFM2-Audio-1.5B-Q8_0.gguf --mmproj $CKPT/mmproj-audioencoder-LFM2-Audio-1.5B-Q8_0.gguf -mv $CKPT/audiodecoder-LFM2-Audio-1.5B-Q8_0.gguf -sys "Perform ASR." --audio $INPUT_WAV
82
  ```
83
 
84
  ### TTS
85
 
86
  TTS requires `-sys "Perform TTS."`, `-p "What is this obsession people have with books?"` for input, and `--output output.wav` for output. It will save audio to `output.wav`.
87
  ```console
88
+ lfm2-audio-<platform>/llama-lfm2-audio -m $CKPT/LFM2-Audio-1.5B-Q8_0.gguf --mmproj $CKPT/mmproj-audioencoder-LFM2-Audio-1.5B-Q8_0.gguf -mv $CKPT/audiodecoder-LFM2-Audio-1.5B-Q8_0.gguf -sys "Perform TTS." -p "What is this obsession people have with books?" --output $OUTPUT_WAV
89
  ```
90
 
91
  ### Interleaved
 
93
  Interleaved produces both, text and audio as output, and can consume text or audio as input.
94
 
95
  ```console
96
+ lfm2-audio-<platform>/llama-lfm2-audio -m $CKPT/LFM2-Audio-1.5B-Q8_0.gguf --mmproj $CKPT/mmproj-audioencoder-LFM2-Audio-1.5B-Q8_0.gguf -mv $CKPT/audiodecoder-LFM2-Audio-1.5B-Q8_0.gguf -sys "Respond with interleaved text and audio." --audio $INPUT_WAV --output $OUTPUT_WAV
97
  ```
98
 
99
  ## Run ASR using `llama-mtmd-cli`
 
101
  Build `llama-mtmd-cli` following the standard build procedure.
102
 
103
  ```console
104
+ lfm2-audio-<platform>/llama-mtmd-cli -m $CKPT/LFM2-Audio-1.5B-Q8_0.gguf --mmproj $CKPT/mmproj-audioencoder-LFM2-Audio-1.5B-Q8_0.gguf -p "<__media__>" -sys "Perform ASR." --audio $INPUT_WAV
105
  ```
106
 
107
  ### Debug
runners/android-arm64/lfm2-audio-android-arm64.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:049be66dee11972d54c95dbe0af6ae802efe98d6bb7647e5267b31e54ded29d9
3
- size 27106971
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:adf22f0ea2598adf9e686720d33a7ad180368b272e297d1186e72d545a547c80
3
+ size 4159454
runners/macos-arm64/lfm2-audio-macos-arm64.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:65d1157ed0054dbf8de737b71970d39db4eafd65d26770d2c1379737ef0ffbbe
3
- size 3841585
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3cd724550d2518c33d5a0bf6c452da921299a0420f07ede80fb06695d541786b
3
+ size 3462697
runners/ubuntu-arm64/lfm2-audio-ubuntu-arm64.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3bf2cf3b448803d5e1a6c54bf3019f0a19f4107aa4b0baf390ed80039bed8a61
3
- size 5708309
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:53a14a2b2c4fc480f202100dda9fdf8268f88b8e655dbf688396ca3e4fdea925
3
+ size 5232143
runners/ubuntu-x64/lfm2-audio-ubuntu-x64.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:071a8affe161f5c953aefc744d56b20772c9041e5ae8bf750e09ed5e5942b7b4
3
- size 6629779
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8f76116e2a4010de10fc5a61a0da5f77020a465478e18ac7fcdabde8e47ebe05
3
+ size 6282015