natalie-a-1
commited on
Commit
Β·
dba3c92
1
Parent(s):
10228ed
Update README.md to enhance model card for Biden Mistral Adapter with detailed descriptions, technical specifications, and usage guidelines.
Browse files
README.md
CHANGED
@@ -1,202 +1,173 @@
|
|
1 |
---
|
2 |
base_model: mistralai/Mistral-7B-Instruct-v0.2
|
3 |
library_name: peft
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
4 |
---
|
5 |
|
6 |
-
|
7 |
|
8 |
-
|
9 |
|
|
|
10 |
|
|
|
11 |
|
12 |
-
|
13 |
-
|
14 |
-
### Model Description
|
15 |
-
|
16 |
-
<!-- Provide a longer summary of what this model is. -->
|
17 |
-
|
18 |
-
|
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 |
-
|
138 |
|
139 |
-
[
|
140 |
|
141 |
-
|
142 |
|
143 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
144 |
|
145 |
-
|
146 |
|
147 |
-
|
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 |
-
|
154 |
|
155 |
-
|
|
|
156 |
|
157 |
-
|
158 |
|
159 |
-
|
160 |
|
161 |
-
|
162 |
|
163 |
-
|
|
|
|
|
164 |
|
165 |
-
|
166 |
|
167 |
-
|
168 |
|
169 |
-
|
170 |
|
171 |
-
|
|
|
172 |
|
173 |
-
|
174 |
|
175 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
176 |
|
177 |
-
|
178 |
|
179 |
-
|
180 |
|
181 |
-
|
|
|
|
|
|
|
182 |
|
183 |
-
|
184 |
|
185 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
186 |
|
187 |
-
|
188 |
|
189 |
-
##
|
190 |
|
191 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
192 |
|
193 |
-
|
194 |
|
195 |
-
|
196 |
|
197 |
-
|
198 |
|
199 |
-
|
200 |
-
### Framework versions
|
201 |
|
202 |
-
|
|
|
|
|
|
1 |
---
|
2 |
base_model: mistralai/Mistral-7B-Instruct-v0.2
|
3 |
library_name: peft
|
4 |
+
language:
|
5 |
+
- en
|
6 |
+
tags:
|
7 |
+
- mistral
|
8 |
+
- lora
|
9 |
+
- adapter
|
10 |
+
- fine-tuned
|
11 |
+
- politics
|
12 |
+
- conversational
|
13 |
+
license: mit
|
14 |
+
datasets:
|
15 |
+
- rohanrao/joe-biden-tweets
|
16 |
+
- christianlillelund/joe-biden-2020-dnc-speech
|
17 |
---
|
18 |
|
19 |
+
<div align="center">
|
20 |
|
21 |
+
# πΊπΈ Biden Mistral Adapter πΊπΈ
|
22 |
|
23 |
+
**A finely-tuned Mistral adapter that captures the distinctive voice, speaking style, and rhetoric of President Joe Biden**
|
24 |
|
25 |
+
</div>
|
26 |
|
27 |
+
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
28 |
|
29 |
+
## π Model Overview
|
30 |
|
31 |
+
This LoRA adapter for [Mistral-7B-Instruct-v0.2](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2) has been crafted to emulate Joe Biden's unique communication patterns, discourse style, and policy framing.
|
32 |
|
33 |
+
### βοΈ Technical Specifications
|
34 |
|
35 |
+
| **Feature** | **Details** |
|
36 |
+
|:---------------:|:--------------------------------------------------:|
|
37 |
+
| Base Model | mistralai/Mistral-7B-Instruct-v0.2 |
|
38 |
+
| Adapter Type | LoRA (Low-Rank Adaptation) |
|
39 |
+
| LoRA Rank | 16 |
|
40 |
+
| Primary Language| English |
|
41 |
+
| Special Feature | Merged weights from style and identity LoRA adapters|
|
42 |
|
43 |
+
---
|
44 |
|
45 |
+
## π Model Evolution
|
|
|
|
|
|
|
|
|
46 |
|
47 |
+
This enhanced version combines two carefully balanced adapter components:
|
48 |
|
49 |
+
- π **Style Adapter**: The original Biden-style adapter (nnat03/biden-mistral-adapter)
|
50 |
+
- π§ **Identity Adapter**: Custom identity adapter (biden-identity-adapter)
|
51 |
|
52 |
+
This merged approach produces more coherent and contextually appropriate responses while preserving Biden's characteristic voice, rhetorical patterns, and speaking cadence.
|
53 |
|
54 |
+
---
|
55 |
|
56 |
+
## π― Intended Applications
|
57 |
|
58 |
+
- π **Educational**: Research on political discourse and communication styles
|
59 |
+
- π **Analytical**: Interactive simulations for rhetoric analysis
|
60 |
+
- π¨ **Creative**: Content development exploring political communication
|
61 |
|
62 |
+
---
|
63 |
|
64 |
+
## π Training Data
|
65 |
|
66 |
+
The model was fine-tuned on two comprehensive datasets:
|
67 |
|
68 |
+
- π± **Biden Twitter Archive** ([2007-2020](https://www.kaggle.com/datasets/rohanrao/joe-biden-tweets)): Capturing everyday communication style
|
69 |
+
- π€ **DNC Acceptance Speech** ([2020](https://www.kaggle.com/datasets/christianlillelund/joe-biden-2020-dnc-speech)): Formal oratorical patterns
|
70 |
|
71 |
+
---
|
72 |
|
73 |
+
## π οΈ Training Process
|
74 |
+
|
75 |
+
### π§ Technical Framework
|
76 |
+
- **Libraries**: Hugging Face Transformers and PEFT
|
77 |
+
- **Optimization**: 4-bit quantization for efficiency
|
78 |
+
|
79 |
+
### π LoRA Configuration
|
80 |
+
```
|
81 |
+
r=16
|
82 |
+
lora_alpha=64
|
83 |
+
lora_dropout=0.05
|
84 |
+
```
|
85 |
+
|
86 |
+
### ποΈ Target Modules
|
87 |
+
- q_proj, k_proj, v_proj, o_proj
|
88 |
+
- gate_proj, up_proj, down_proj
|
89 |
+
|
90 |
+
### βοΈ Training Parameters
|
91 |
+
| **Parameter** | **Value** |
|
92 |
+
|:------------:|:---------:|
|
93 |
+
| Batch Size | 4 |
|
94 |
+
| Gradient Accumulation | 4 |
|
95 |
+
| Learning Rate | 2e-4 |
|
96 |
+
| Epochs | 3 |
|
97 |
+
| LR Scheduler | cosine |
|
98 |
+
| Optimizer | paged_adamw_8bit |
|
99 |
+
| Precision | BF16 |
|
100 |
|
101 |
+
---
|
102 |
|
103 |
+
## β οΈ Limitations and Considerations
|
104 |
|
105 |
+
- π This model mimics a speaking style but may not provide factually accurate information
|
106 |
+
- π£οΈ While it emulates Biden's rhetoric, it does not represent his actual views
|
107 |
+
- π The model may reproduce biases present in the training data
|
108 |
+
- π Not suitable for production applications without RAG enhancement for factual accuracy
|
109 |
|
110 |
+
---
|
111 |
|
112 |
+
## π» Implementation Guide
|
113 |
+
|
114 |
+
```python
|
115 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig
|
116 |
+
from peft import PeftModel
|
117 |
+
import torch
|
118 |
+
|
119 |
+
# Load base model with 4-bit quantization
|
120 |
+
base_model_id = "mistralai/Mistral-7B-Instruct-v0.2"
|
121 |
+
bnb_config = BitsAndBytesConfig(
|
122 |
+
load_in_4bit=True,
|
123 |
+
bnb_4bit_compute_dtype=torch.float16,
|
124 |
+
bnb_4bit_quant_type="nf4",
|
125 |
+
bnb_4bit_use_double_quant=True,
|
126 |
+
)
|
127 |
+
|
128 |
+
# Load model and tokenizer
|
129 |
+
model = AutoModelForCausalLM.from_pretrained(
|
130 |
+
base_model_id,
|
131 |
+
quantization_config=bnb_config,
|
132 |
+
device_map="auto",
|
133 |
+
torch_dtype=torch.float16
|
134 |
+
)
|
135 |
+
tokenizer = AutoTokenizer.from_pretrained(base_model_id)
|
136 |
+
|
137 |
+
# Apply the adapter
|
138 |
+
model = PeftModel.from_pretrained(model, "nnat03/biden-mistral-adapter")
|
139 |
+
|
140 |
+
# Generate a response
|
141 |
+
prompt = "What's your vision for America's future?"
|
142 |
+
input_text = f"<s>[INST] {prompt} [/INST]"
|
143 |
+
inputs = tokenizer(input_text, return_tensors="pt").to("cuda")
|
144 |
+
outputs = model.generate(**inputs, max_length=512, temperature=0.7, do_sample=True)
|
145 |
+
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
|
146 |
+
print(response.split("[/INST]")[-1].strip())
|
147 |
+
```
|
148 |
|
149 |
+
---
|
150 |
|
151 |
+
## π Citation
|
152 |
|
153 |
+
```bibtex
|
154 |
+
@misc{nnat03-biden-mistral-adapter,
|
155 |
+
author = {nnat03},
|
156 |
+
title = {Biden Mistral Adapter},
|
157 |
+
year = {2023},
|
158 |
+
publisher = {Hugging Face},
|
159 |
+
howpublished = {\url{https://huggingface.co/nnat03/biden-mistral-adapter}}
|
160 |
+
}
|
161 |
+
```
|
162 |
|
163 |
+
---
|
164 |
|
165 |
+
## π Ethical Usage
|
166 |
|
167 |
+
This model is intended for educational and research purposes only. It mimics the speaking style of a public figure but does not represent their actual views or statements. Please use responsibly.
|
168 |
|
169 |
+
---
|
|
|
170 |
|
171 |
+
<div align="center">
|
172 |
+
<b>Framework version:</b> PEFT 0.15.0
|
173 |
+
</div>
|