Remove files from LFS tracking and re-add as regular text files
Browse files
README.md
CHANGED
|
@@ -1,3 +1,167 @@
|
|
| 1 |
-
|
| 2 |
-
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: proprietary
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
+
- zh
|
| 6 |
+
library_name: transformers
|
| 7 |
+
pipeline_tag: text-generation
|
| 8 |
+
tags:
|
| 9 |
+
- reasoning
|
| 10 |
+
- mathematics
|
| 11 |
+
- programming
|
| 12 |
+
- creative-writing
|
| 13 |
+
- chain-of-thought
|
| 14 |
+
---
|
| 15 |
+
|
| 16 |
+
# Brello Thinking
|
| 17 |
+
|
| 18 |
+
<p align="center">
|
| 19 |
+
<img src="https://via.placeholder.com/400x200/4A90E2/FFFFFF?text=BRELLO+THINKING" width="400"/> <br>
|
| 20 |
+
</p>
|
| 21 |
+
|
| 22 |
+
<p align="center">
|
| 23 |
+
π€ <a href="#"><b>Epic Systems</b></a> |
|
| 24 |
+
π¨βπ» <a href="#"><b>Rehan Temkar</b></a> |
|
| 25 |
+
π§ <a href="#"><b>Advanced Reasoning</b></a>
|
| 26 |
+
</p>
|
| 27 |
+
|
| 28 |
+
## Model Introduction
|
| 29 |
+
|
| 30 |
+
**Brello Thinking** is an advanced language model created by **Epic Systems** and engineered by **Rehan Temkar**. Built on the robust Tencent Hunyuan base model, Brello Thinking specializes in deep reasoning, mathematical problem-solving, coding, and creative thinking with enhanced chain-of-thought capabilities.
|
| 31 |
+
|
| 32 |
+
### Key Features and Advantages
|
| 33 |
+
|
| 34 |
+
- **π§ Advanced Reasoning**: Enhanced chain-of-thought capabilities with both fast and slow thinking modes
|
| 35 |
+
- **π’ Mathematical Excellence**: Superior performance in mathematical problem-solving and computation
|
| 36 |
+
- **π» Programming Prowess**: Strong coding capabilities across multiple programming languages
|
| 37 |
+
- **π Long Context Understanding**: Supports extended conversations and document analysis
|
| 38 |
+
- **π― Creative Problem Solving**: Innovative approaches to complex problems
|
| 39 |
+
- **π Multi-language Support**: Fluent in multiple languages with cultural understanding
|
| 40 |
+
|
| 41 |
+
## Model Architecture
|
| 42 |
+
|
| 43 |
+
- **Base Model**: Tencent Hunyuan
|
| 44 |
+
- **Parameters**: 1.8B (optimized for efficiency)
|
| 45 |
+
- **Context Window**: 256K tokens
|
| 46 |
+
- **Architecture**: HunYuanDenseV1ForCausalLM
|
| 47 |
+
- **Specialization**: Reasoning, Mathematics, Programming, Creative Thinking
|
| 48 |
+
|
| 49 |
+
## Usage
|
| 50 |
+
|
| 51 |
+
### Basic Usage
|
| 52 |
+
|
| 53 |
+
```python
|
| 54 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer
|
| 55 |
+
|
| 56 |
+
# Load Brello Thinking
|
| 57 |
+
model_name = "BrelloES/brello-thinking"
|
| 58 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
| 59 |
+
model = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto")
|
| 60 |
+
|
| 61 |
+
# Example conversation
|
| 62 |
+
messages = [
|
| 63 |
+
{"role": "user", "content": "What is 2+2?"}
|
| 64 |
+
]
|
| 65 |
+
|
| 66 |
+
tokenized_chat = tokenizer.apply_chat_template(
|
| 67 |
+
messages,
|
| 68 |
+
tokenize=True,
|
| 69 |
+
add_generation_prompt=True,
|
| 70 |
+
return_tensors="pt",
|
| 71 |
+
enable_thinking=True
|
| 72 |
+
)
|
| 73 |
+
|
| 74 |
+
outputs = model.generate(
|
| 75 |
+
tokenized_chat.to(model.device),
|
| 76 |
+
max_new_tokens=2048,
|
| 77 |
+
do_sample=True,
|
| 78 |
+
top_k=20,
|
| 79 |
+
top_p=0.8,
|
| 80 |
+
repetition_penalty=1.05,
|
| 81 |
+
temperature=0.7
|
| 82 |
+
)
|
| 83 |
+
|
| 84 |
+
response = tokenizer.decode(outputs[0])
|
| 85 |
+
print(response)
|
| 86 |
+
```
|
| 87 |
+
|
| 88 |
+
### Thinking Mode
|
| 89 |
+
|
| 90 |
+
Brello Thinking supports enhanced reasoning with thinking mode:
|
| 91 |
+
|
| 92 |
+
```python
|
| 93 |
+
# Enable thinking mode (default)
|
| 94 |
+
tokenized_chat = tokenizer.apply_chat_template(
|
| 95 |
+
messages,
|
| 96 |
+
enable_thinking=True # Shows reasoning process
|
| 97 |
+
)
|
| 98 |
+
|
| 99 |
+
# Disable thinking mode
|
| 100 |
+
tokenized_chat = tokenizer.apply_chat_template(
|
| 101 |
+
messages,
|
| 102 |
+
enable_thinking=False # Direct answers
|
| 103 |
+
)
|
| 104 |
+
```
|
| 105 |
+
|
| 106 |
+
## Model Capabilities
|
| 107 |
+
|
| 108 |
+
### Mathematical Reasoning
|
| 109 |
+
- Complex mathematical problem-solving
|
| 110 |
+
- Step-by-step mathematical proofs
|
| 111 |
+
- Statistical analysis and computation
|
| 112 |
+
|
| 113 |
+
### Programming
|
| 114 |
+
- Code generation in multiple languages
|
| 115 |
+
- Debugging and code optimization
|
| 116 |
+
- Algorithm design and implementation
|
| 117 |
+
|
| 118 |
+
### Creative Writing
|
| 119 |
+
- Story generation and creative content
|
| 120 |
+
- Technical writing and documentation
|
| 121 |
+
- Poetry and artistic expression
|
| 122 |
+
|
| 123 |
+
### Problem Solving
|
| 124 |
+
- Logical reasoning and analysis
|
| 125 |
+
- Critical thinking and evaluation
|
| 126 |
+
- Strategic planning and decision-making
|
| 127 |
+
|
| 128 |
+
## Technical Specifications
|
| 129 |
+
|
| 130 |
+
| Specification | Value |
|
| 131 |
+
|---------------|-------|
|
| 132 |
+
| Model Size | 1.8B Parameters |
|
| 133 |
+
| Context Window | 256K Tokens |
|
| 134 |
+
| Architecture | HunYuanDenseV1ForCausalLM |
|
| 135 |
+
| Base Model | Tencent Hunyuan |
|
| 136 |
+
| Creator | Epic Systems |
|
| 137 |
+
| Engineer | Rehan Temkar |
|
| 138 |
+
| License | Proprietary - Epic Systems |
|
| 139 |
+
|
| 140 |
+
## Performance
|
| 141 |
+
|
| 142 |
+
Brello Thinking demonstrates superior performance in:
|
| 143 |
+
- Mathematical reasoning and computation
|
| 144 |
+
- Programming and code generation
|
| 145 |
+
- Creative problem-solving
|
| 146 |
+
- Long-context understanding
|
| 147 |
+
- Multi-language tasks
|
| 148 |
+
|
| 149 |
+
## License
|
| 150 |
+
|
| 151 |
+
This model is proprietary software created by Epic Systems and engineered by Rehan Temkar. All rights reserved.
|
| 152 |
+
|
| 153 |
+
## Contact
|
| 154 |
+
|
| 155 |
+
- **Creator**: Epic Systems
|
| 156 |
+
- **Engineer**: Rehan Temkar
|
| 157 |
+
- **Model**: Brello Thinking v1.0.0
|
| 158 |
+
|
| 159 |
+
---
|
| 160 |
+
|
| 161 |
+
*Brello Thinking - Advanced AI Reasoning by Epic Systems*
|
| 162 |
+
|
| 163 |
+
---
|
| 164 |
+
|
| 165 |
+
## Original Model Information
|
| 166 |
+
|
| 167 |
+
This model is based on the Tencent Hunyuan model. For more information about the original model, see the [Hunyuan documentation](https://github.com/Tencent-Hunyuan/Hunyuan-1.8B).
|