farzadab commited on
Commit
375c671
1 Parent(s): a716fe2

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +68 -154
README.md CHANGED
@@ -1,199 +1,113 @@
1
  ---
 
 
 
 
 
 
 
 
 
 
 
2
  library_name: transformers
3
- tags: []
 
 
 
 
 
 
4
  ---
5
 
6
- # Model Card for Model ID
7
 
8
- <!-- Provide a quick summary of what the model is/does. -->
9
 
 
10
 
11
 
12
  ## Model Details
13
 
14
  ### Model Description
15
 
16
- <!-- Provide a longer summary of what this model is. -->
 
 
17
 
18
- This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
 
19
 
20
- - **Developed by:** [More Information Needed]
21
- - **Funded by [optional]:** [More Information Needed]
22
- - **Shared by [optional]:** [More Information Needed]
23
- - **Model type:** [More Information Needed]
24
- - **Language(s) (NLP):** [More Information Needed]
25
- - **License:** [More Information Needed]
26
- - **Finetuned from model [optional]:** [More Information Needed]
27
 
28
- ### Model Sources [optional]
29
 
30
- <!-- Provide the basic links for the model. -->
 
31
 
32
- - **Repository:** [More Information Needed]
33
- - **Paper [optional]:** [More Information Needed]
34
- - **Demo [optional]:** [More Information Needed]
35
 
36
- ## Uses
37
 
38
- <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
 
 
39
 
40
- ### Direct Use
 
 
41
 
42
- <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
43
 
44
- [More Information Needed]
 
45
 
46
- ### Downstream Use [optional]
47
 
48
- <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
 
 
 
 
 
 
 
49
 
50
- [More Information Needed]
51
 
52
- ### Out-of-Scope Use
53
-
54
- <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
55
-
56
- [More Information Needed]
57
-
58
- ## Bias, Risks, and Limitations
59
-
60
- <!-- This section is meant to convey both technical and sociotechnical limitations. -->
61
-
62
- [More Information Needed]
63
-
64
- ### Recommendations
65
-
66
- <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
67
-
68
- Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
69
-
70
- ## How to Get Started with the Model
71
 
72
- Use the code below to get started with the model.
73
 
74
- [More Information Needed]
75
 
76
- ## Training Details
77
 
78
  ### Training Data
79
 
80
- <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
81
 
82
- [More Information Needed]
83
 
84
  ### Training Procedure
85
 
86
- <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
87
-
88
- #### Preprocessing [optional]
89
-
90
- [More Information Needed]
91
 
92
 
93
  #### Training Hyperparameters
94
 
95
- - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
 
96
 
97
- #### Speeds, Sizes, Times [optional]
98
 
99
- <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
100
 
101
- [More Information Needed]
102
 
103
  ## Evaluation
104
-
105
- <!-- This section describes the evaluation protocols and provides the results. -->
106
-
107
- ### Testing Data, Factors & Metrics
108
-
109
- #### Testing Data
110
-
111
- <!-- This should link to a Dataset Card if possible. -->
112
-
113
- [More Information Needed]
114
-
115
- #### Factors
116
-
117
- <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
118
-
119
- [More Information Needed]
120
-
121
- #### Metrics
122
-
123
- <!-- These are the evaluation metrics being used, ideally with a description of why. -->
124
-
125
- [More Information Needed]
126
-
127
- ### Results
128
-
129
- [More Information Needed]
130
-
131
- #### Summary
132
-
133
-
134
-
135
- ## Model Examination [optional]
136
-
137
- <!-- Relevant interpretability work for the model goes here -->
138
-
139
- [More Information Needed]
140
-
141
- ## Environmental Impact
142
-
143
- <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
144
-
145
- Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
146
-
147
- - **Hardware Type:** [More Information Needed]
148
- - **Hours used:** [More Information Needed]
149
- - **Cloud Provider:** [More Information Needed]
150
- - **Compute Region:** [More Information Needed]
151
- - **Carbon Emitted:** [More Information Needed]
152
-
153
- ## Technical Specifications [optional]
154
-
155
- ### Model Architecture and Objective
156
-
157
- [More Information Needed]
158
-
159
- ### Compute Infrastructure
160
-
161
- [More Information Needed]
162
-
163
- #### Hardware
164
-
165
- [More Information Needed]
166
-
167
- #### Software
168
-
169
- [More Information Needed]
170
-
171
- ## Citation [optional]
172
-
173
- <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
174
-
175
- **BibTeX:**
176
-
177
- [More Information Needed]
178
-
179
- **APA:**
180
-
181
- [More Information Needed]
182
-
183
- ## Glossary [optional]
184
-
185
- <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
186
-
187
- [More Information Needed]
188
-
189
- ## More Information [optional]
190
-
191
- [More Information Needed]
192
-
193
- ## Model Card Authors [optional]
194
-
195
- [More Information Needed]
196
-
197
- ## Model Card Contact
198
-
199
- [More Information Needed]
 
1
  ---
2
+ language:
3
+ - en
4
+ - ar
5
+ - de
6
+ - es
7
+ - fr
8
+ - it
9
+ - ja
10
+ - pt
11
+ - ru
12
+ license: mit
13
  library_name: transformers
14
+ datasets:
15
+ - fixie-ai/librispeech_asr
16
+ - fixie-ai/common_voice_17_0
17
+ - fixie-ai/peoples_speech
18
+ - fnlp/AnyInstruct
19
+ metrics:
20
+ - bleu
21
  ---
22
 
23
+ # Model Card for Ultravox
24
 
25
+ Ultravox is a multimodal Speech LLM built around a pretrained [Llama3.1-70B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-70B) and [Whisper-medium](https://huggingface.co/openai/whisper-medium) backbone.
26
 
27
+ See https://ultravox.ai for the GitHub repo and more information.
28
 
29
 
30
  ## Model Details
31
 
32
  ### Model Description
33
 
34
+ Ultravox is a multimodal model that can consume both speech and text as input (e.g., a text system prompt and voice user message).
35
+ The input to the model is given as a text prompt with a special `<|audio|>` pseudo-token, and the model processor will replace this magic token with embeddings derived from the input audio.
36
+ Using the merged embeddings as input, the model will then generate output text as usual.
37
 
38
+ In a future revision of Ultravox, we plan to expand the token vocabulary to support generation of semantic and acoustic audio tokens, which can then be fed to a vocoder to produce voice output.
39
+ No preference tuning has been applied to this revision of the model.
40
 
41
+ - **Developed by:** Fixie.ai
42
+ - **License:** MIT
 
 
 
 
 
43
 
44
+ ### Model Sources
45
 
46
+ - **Repository:** https://ultravox.ai
47
+ - **Demo:** See repo
48
 
49
+ ## Usage
 
 
50
 
51
+ Think of the model as an LLM that can also hear and understand speech. As such, it can be used as a voice agent, and also to do speech-to-speech translation, analysis of spoken audio, etc.
52
 
53
+ To use the model, try the following:
54
+ ```python
55
+ # pip install transformers peft librosa
56
 
57
+ import transformers
58
+ import numpy as np
59
+ import librosa
60
 
61
+ pipe = transformers.pipeline(model='fixie-ai/ultravox-v0_4', trust_remote_code=True)
62
 
63
+ path = "<path-to-input-audio>" # TODO: pass the audio here
64
+ audio, sr = librosa.load(path, sr=16000)
65
 
 
66
 
67
+ turns = [
68
+ {
69
+ "role": "system",
70
+ "content": "You are a friendly and helpful character. You love to answer questions for people."
71
+ },
72
+ ]
73
+ pipe({'audio': audio, 'turns': turns, 'sampling_rate': sr}, max_new_tokens=30)
74
+ ```
75
 
 
76
 
77
+ ## Training Details
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
78
 
79
+ The model uses a pre-trained [Llama3.1-70B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-70B) backbone as well as the encoder part of [Whisper-medium](https://huggingface.co/openai/whisper-medium).
80
 
81
+ Only the multi-modal adapter is trained, while Whisper encoder and Llama are kept frozen.
82
 
83
+ We use a knowledge-distillation loss where Ultravox is trying to match the logits of the text-based Llama backbone.
84
 
85
  ### Training Data
86
 
87
+ Training dataset is a mix of ASR datasets, extended by adding a "continuation" generated by Llama 3.1 70B.
88
 
 
89
 
90
  ### Training Procedure
91
 
92
+ Supervised speech to audio finetuning. For more info, see [training code in Ultravox repo](https://github.com/fixie-ai/ultravox/blob/main/ultravox/training/train.py).
 
 
 
 
93
 
94
 
95
  #### Training Hyperparameters
96
 
97
+ - **Training regime:** BF16 mixed precision training
98
+ - **Hardward used:** 8x H100 GPUs
99
 
100
+ #### Speeds, Sizes, Times
101
 
102
+ The current version of Ultravox, when invoked with audio content, has a time-to-first-token (TTFT) of approximately 400ms, and a tokens-per-second rate of ~50-100 when using 4xA100-40GB GPU, all using a Llama 3.1 70B backbone.
103
 
104
+ Check out the audio tab on [TheFastest.ai](https://thefastest.ai/?m=audio) for daily benchmarks and a comparison with other existing models.
105
 
106
  ## Evaluation
107
+ | | en_de (BLEU) | es_en (BLEU) | LibriSpeech clean.test (WER) |
108
+ |:------------------|:-------------|:-------------|:----------------------------|
109
+ | Ultravox v0.3 | 22.66 | 24.74 | 6.67 |
110
+ | Ultravox v0.4 8B | 25.47 | 37.11 | 4.45 |
111
+ | **Ultravox v0.4 70B** | **30.30** | **39.55** | **4.49** |
112
+ | Llama3.1 8B (text-only) | 31.95 | 44.62 | - |
113
+ | Llama3.1 70B (text-only) | 32.59 | 46.39 | - |