Hugging Face's logo

pranav2711
/
ncu-smartllm-fp32

Safetensors
mistral
8-bit precision
Model card Files
xet
Community
pranav2711 commited on
Commit
15da20e
·
verified ·
1 Parent(s): ffc6e20

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +19 -27
README.md CHANGED
@@ -4,7 +4,7 @@ base_model:
4
  - mistralai/Mistral-7B-Instruct-v0.3
5
  ---
6
 
7
- # 🧠 NCU SmartLLM (FP32) — Fine-Tuned Mistral-7B
8
 
9
  [![Downloads](https://img.shields.io/badge/downloads-53-blue)](https://huggingface.co/pranav2711/ncu-smartllm-fp32)
10
  [![License: Apache-2.0](https://img.shields.io/badge/license-Apache%202.0-green.svg)](LICENSE)
@@ -13,7 +13,7 @@ base_model:
13
 
14
  ---
15
 
16
- ## 📌 Model Summary
17
 
18
  - **Base Model:** [Mistral-7B](https://huggingface.co/mistralai/Mistral-7B-v0.1)
19
  - **Fine-tuning:** Full fine-tuning (FP32)
@@ -24,7 +24,7 @@ base_model:
24
 
25
  ---
26
 
27
- ## 🧠 Capabilities
28
 
29
  The model has been trained to:
30
  - Answer FAQs related to NCU (hostel, fees, scholarships, re-evaluation, etc.)
@@ -32,13 +32,13 @@ The model has been trained to:
32
  - Handle general admin-related queries
33
  - Demonstrate Mistral's capabilities in an Indian academic context
34
 
35
- > ✅ Fully plug-and-play with Hugging Face Transformers & inference APIs
36
 
37
  ---
38
 
39
- ## 💻 Usage
40
 
41
- ### ▶️ Inference (Transformers)
42
 
43
  ```bash
44
  pip install transformers accelerate
@@ -58,28 +58,20 @@ print(tokenizer.decode(outputs[0], skip_special_tokens=True))
58
 
59
  ---
60
 
61
- ## 🚀 Try It Online
62
-
63
- You can try the model on Hugging Face Spaces:
64
-
65
- 🔗 [Launch NCU Smart Chat (Gradio UI)](https://huggingface.co/spaces/pranav2711/phi2-ncu-chat)
66
-
67
- ---
68
-
69
- ## 🧠 Model Training Details
70
 
71
  | Detail | Value |
72
  | ------------------ | ------------------------- |
73
  | Base model | mistralai/Mistral-7B-v0.1 |
74
- | Fine-tuned epochs | 4 |
75
- | Batch size | 1–2 |
76
  | Tokenizer max len | 512 |
77
  | Output format | Instruction → Response |
78
- | Loss (final epoch) | \~2.1 |
79
 
80
  ---
81
 
82
- ## 🧾 Dataset Format (Instruction-Tuned)
83
 
84
  Each entry follows:
85
 
@@ -103,7 +95,7 @@ Formatted as:
103
 
104
  ---
105
 
106
- ## 🧠 Model Applications
107
 
108
  * College ERP chatbots
109
  * Student helpdesks
@@ -112,23 +104,23 @@ Formatted as:
112
 
113
  ---
114
 
115
- ## 📃 License
116
 
117
  This model is released under the **Apache 2.0 license**. You are free to use, modify, and distribute it with attribution.
118
 
119
  ---
120
 
121
- ## 👤 Author & Maintainer
122
 
123
- * 👨‍💻 [Pranav Singh](https://huggingface.co/pranav2711)
124
- * 🌐 The NorthCap University (NCU), Gurugram
125
- * ✉️ For collaboration: ping on Hugging Face or GitHub
126
 
127
  ---
128
 
129
- ## ⭐ How to Support
130
 
131
- * Star ⭐ this model on Hugging Face
132
  * Try it on Spaces and share your feedback
133
  * Contribute improvements or suggest datasets
134
 
 
4
  - mistralai/Mistral-7B-Instruct-v0.3
5
  ---
6
 
7
+ # NCU SmartLLM (FP32) — Fine-Tuned Mistral-7B
8
 
9
  [![Downloads](https://img.shields.io/badge/downloads-53-blue)](https://huggingface.co/pranav2711/ncu-smartllm-fp32)
10
  [![License: Apache-2.0](https://img.shields.io/badge/license-Apache%202.0-green.svg)](LICENSE)
 
13
 
14
  ---
15
 
16
+ ## Model Summary
17
 
18
  - **Base Model:** [Mistral-7B](https://huggingface.co/mistralai/Mistral-7B-v0.1)
19
  - **Fine-tuning:** Full fine-tuning (FP32)
 
24
 
25
  ---
26
 
27
+ ## Capabilities
28
 
29
  The model has been trained to:
30
  - Answer FAQs related to NCU (hostel, fees, scholarships, re-evaluation, etc.)
 
32
  - Handle general admin-related queries
33
  - Demonstrate Mistral's capabilities in an Indian academic context
34
 
35
+ > Fully plug-and-play with Hugging Face Transformers & inference APIs
36
 
37
  ---
38
 
39
+ ## Usage
40
 
41
+ ### Inference (Transformers)
42
 
43
  ```bash
44
  pip install transformers accelerate
 
58
 
59
  ---
60
 
61
+ ## Model Training Details
 
 
 
 
 
 
 
 
62
 
63
  | Detail | Value |
64
  | ------------------ | ------------------------- |
65
  | Base model | mistralai/Mistral-7B-v0.1 |
66
+ | Fine-tuned epochs | 10 |
67
+ | Batch size | 2 |
68
  | Tokenizer max len | 512 |
69
  | Output format | Instruction → Response |
70
+ | Loss (final epoch) | \~1.405300 |
71
 
72
  ---
73
 
74
+ ## Dataset Format (Instruction-Tuned)
75
 
76
  Each entry follows:
77
 
 
95
 
96
  ---
97
 
98
+ ## Model Applications
99
 
100
  * College ERP chatbots
101
  * Student helpdesks
 
104
 
105
  ---
106
 
107
+ ## License
108
 
109
  This model is released under the **Apache 2.0 license**. You are free to use, modify, and distribute it with attribution.
110
 
111
  ---
112
 
113
+ ## Author & Maintainer
114
 
115
+ * [Pranav Singh](https://huggingface.co/pranav2711)
116
+ * The NorthCap University (NCU), Gurugram
117
+ * For collaboration: ping on Hugging Face or GitHub
118
 
119
  ---
120
 
121
+ ## How to Support
122
 
123
+ * Star this model on Hugging Face
124
  * Try it on Spaces and share your feedback
125
  * Contribute improvements or suggest datasets
126