File size: 7,953 Bytes
b5b75b0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
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
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
---
base_model:
- Undi95/Llama-3-Unholy-8B
- Locutusque/llama-3-neural-chat-v1-8b
- ruslanmv/Medical-Llama3-8B-16bit
library_name: transformers
tags:
- mergekit
- merge
- medical
license: other
datasets:
- mlabonne/orpo-dpo-mix-40k
- Open-Orca/SlimOrca-Dedup
- jondurbin/airoboros-3.2
- microsoft/orca-math-word-problems-200k
- m-a-p/Code-Feedback
- MaziyarPanahi/WizardLM_evol_instruct_V2_196k
- ruslanmv/ai-medical-chatbot
model-index:
- name: Medichat-Llama3-8B
  results:
  - task:
      type: text-generation
      name: Text Generation
    dataset:
      name: AI2 Reasoning Challenge (25-Shot)
      type: ai2_arc
      config: ARC-Challenge
      split: test
      args:
        num_few_shot: 25
    metrics:
    - type: acc_norm
      value: 59.13
      name: normalized accuracy
    source:
      url: >-
        https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=sethuiyer/Medichat-Llama3-8B
      name: Open LLM Leaderboard
  - task:
      type: text-generation
      name: Text Generation
    dataset:
      name: HellaSwag (10-Shot)
      type: hellaswag
      split: validation
      args:
        num_few_shot: 10
    metrics:
    - type: acc_norm
      value: 82.9
      name: normalized accuracy
    source:
      url: >-
        https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=sethuiyer/Medichat-Llama3-8B
      name: Open LLM Leaderboard
  - task:
      type: text-generation
      name: Text Generation
    dataset:
      name: MMLU (5-Shot)
      type: cais/mmlu
      config: all
      split: test
      args:
        num_few_shot: 5
    metrics:
    - type: acc
      value: 60.35
      name: accuracy
    source:
      url: >-
        https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=sethuiyer/Medichat-Llama3-8B
      name: Open LLM Leaderboard
  - task:
      type: text-generation
      name: Text Generation
    dataset:
      name: TruthfulQA (0-shot)
      type: truthful_qa
      config: multiple_choice
      split: validation
      args:
        num_few_shot: 0
    metrics:
    - type: mc2
      value: 49.65
    source:
      url: >-
        https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=sethuiyer/Medichat-Llama3-8B
      name: Open LLM Leaderboard
  - task:
      type: text-generation
      name: Text Generation
    dataset:
      name: Winogrande (5-shot)
      type: winogrande
      config: winogrande_xl
      split: validation
      args:
        num_few_shot: 5
    metrics:
    - type: acc
      value: 78.93
      name: accuracy
    source:
      url: >-
        https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=sethuiyer/Medichat-Llama3-8B
      name: Open LLM Leaderboard
  - task:
      type: text-generation
      name: Text Generation
    dataset:
      name: GSM8k (5-shot)
      type: gsm8k
      config: main
      split: test
      args:
        num_few_shot: 5
    metrics:
    - type: acc
      value: 60.35
      name: accuracy
    source:
      url: >-
        https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=sethuiyer/Medichat-Llama3-8B
      name: Open LLM Leaderboard
language:
- en
---

### Medichat-Llama3-8B

Built upon the powerful LLaMa-3 architecture and fine-tuned on an extensive dataset of health information, this model leverages its vast medical knowledge to offer clear, comprehensive answers.

This model is generally better for accurate and informative responses, particularly for users seeking in-depth medical advice.


The following YAML configuration was used to produce this model:

```yaml

models:
  - model: Undi95/Llama-3-Unholy-8B
    parameters:
      weight: [0.25, 0.35, 0.45, 0.35, 0.25]
      density: [0.1, 0.25, 0.5, 0.25, 0.1]
  - model: Locutusque/llama-3-neural-chat-v1-8b
  - model: ruslanmv/Medical-Llama3-8B-16bit
    parameters:
      weight: [0.55, 0.45, 0.35, 0.45, 0.55]
      density: [0.1, 0.25, 0.5, 0.25, 0.1]
merge_method: dare_ties
base_model: Locutusque/llama-3-neural-chat-v1-8b
parameters:
  int8_mask: true
dtype: bfloat16

```

# Comparision Against Dr.Samantha 7B

| Subject                 | Medichat-Llama3-8B Accuracy (%) | Dr. Samantha Accuracy (%) |
|-------------------------|---------------------------------|---------------------------|
| Clinical Knowledge      | 71.70                           | 52.83                     |
| Medical Genetics        | 78.00                           | 49.00                     |
| Human Aging             | 70.40                           | 58.29                     |
| Human Sexuality         | 73.28                           | 55.73                     |
| College Medicine        | 62.43                           | 38.73                     |
| Anatomy                 | 64.44                           | 41.48                     |
| College Biology         | 72.22                           | 52.08                     |
| High School Biology     | 77.10                           | 53.23                     |
| Professional Medicine   | 63.97                           | 38.73                     |
| Nutrition               | 73.86                           | 50.33                     |
| Professional Psychology | 68.95                           | 46.57                     |
| Virology                | 54.22                           | 41.57                     |
| High School Psychology  | 83.67                           | 66.60                     |
| **Average**             | **70.33**                       | **48.85**                 |


The current model demonstrates a substantial improvement over the previous [Dr. Samantha](sethuiyer/Dr_Samantha-7b) model in terms of subject-specific knowledge and accuracy.

### Usage:
```python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM

class MedicalAssistant:
    def __init__(self, model_name="sethuiyer/Medichat-Llama3-8B", device="cuda"):
        self.device = device
        self.tokenizer = AutoTokenizer.from_pretrained(model_name)
        self.model = AutoModelForCausalLM.from_pretrained(model_name).to(self.device)
        self.sys_message = ''' 
        You are an AI Medical Assistant trained on a vast dataset of health information. Please be thorough and
        provide an informative answer. If you don't know the answer to a specific medical inquiry, advise seeking professional help.
        '''

    def format_prompt(self, question):
        messages = [
            {"role": "system", "content": self.sys_message},
            {"role": "user", "content": question}
        ]
        prompt = self.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
        return prompt

    def generate_response(self, question, max_new_tokens=512):
        prompt = self.format_prompt(question)
        inputs = self.tokenizer(prompt, return_tensors="pt").to(self.device)
        with torch.no_grad():
            outputs = self.model.generate(**inputs, max_new_tokens=max_new_tokens, use_cache=True)
        answer = self.tokenizer.batch_decode(outputs, skip_special_tokens=True)[0].strip()
        return answer

if __name__ == "__main__":
    assistant = MedicalAssistant()
    question = '''
    Symptoms:
    Dizziness, headache, and nausea.

    What is the differential diagnosis?
    '''
    response = assistant.generate_response(question)
    print(response)

```

## Quants
Thanks to [Quant Factory](https://huggingface.co/QuantFactory), the quantized version of this model is available at [QuantFactory/Medichat-Llama3-8B-GGUF](https://huggingface.co/QuantFactory/Medichat-Llama3-8B-GGUF),


## Ollama
This model is now also available on Ollama. You can use it by running the command ```ollama run monotykamary/medichat-llama3``` in your 
terminal. If you have limited computing resources, check out this [video](https://www.youtube.com/watch?v=Qa1h7ygwQq8) to learn how to run it on 
a Google Colab backend.