|
--- |
|
language: |
|
- en |
|
tags: |
|
- falcon3 |
|
--- |
|
|
|
|
|
# Table of Contents |
|
|
|
0. [TL;DR](#TL;DR) |
|
1. [Model Details](#model-details) |
|
2. [Usage](#usage) |
|
3. [Training Details](#training-details) |
|
4. [Evaluation](#evaluation) |
|
|
|
|
|
# TL;DR |
|
|
|
# Model Details |
|
|
|
⚠️ **This is a raw, pretrained model, which should be further finetuned for most usecases.** |
|
|
|
## Model Description |
|
|
|
- **Developed by:** [https://www.tii.ae](https://www.tii.ae) |
|
- **Model type:** Causal decoder-only |
|
- **Architecture:** Transformer-base |
|
- **Language(s) (NLP):** Mainly English |
|
- **License:** TII Falcon-LLM License 2.0 |
|
|
|
<br> |
|
|
|
# Usage |
|
|
|
Find below some example scripts on how to use the model in `transformers` (Make sure to have the latest transformers, or the one built from source): |
|
|
|
## Using the Pytorch model with 🤗 transformers |
|
|
|
### Running the model on a CPU |
|
|
|
<details> |
|
<summary> Click to expand </summary> |
|
|
|
```python |
|
from transformers import AutoTokenizer, AutoModelForCausalLM |
|
|
|
tokenizer = AutoTokenizer.from_pretrained("tiiuae/Falcon3-7B-Base") |
|
model = AutoModelForCausalLM.from_pretrained("tiiuae/Falcon3-7B-Base") |
|
|
|
input_text = "Question: How many hours in one day? Answer: " |
|
input_ids = tokenizer(input_text, return_tensors="pt").input_ids |
|
|
|
outputs = model.generate(input_ids) |
|
print(tokenizer.decode(outputs[0])) |
|
``` |
|
|
|
</details> |
|
|
|
### Running the model on a GPU |
|
|
|
<details> |
|
<summary> Click to expand </summary> |
|
|
|
```python |
|
# pip install accelerate |
|
from transformers import AutoTokenizer, AutoModelForCausalLM |
|
|
|
tokenizer = AutoTokenizer.from_pretrained("tiiuae/Falcon3-7B-Base") |
|
model = AutoModelForCausalLM.from_pretrained("tiiuae/Falcon3-7B-Base", device_map="auto") |
|
|
|
input_text = "Question: How many hours in one day? Answer: " |
|
input_ids = tokenizer(input_text, return_tensors="pt").input_ids.to("cuda") |
|
|
|
outputs = model.generate(input_ids) |
|
print(tokenizer.decode(outputs[0])) |
|
``` |
|
|
|
</details> |
|
|
|
### Running the model on a GPU using `torch.compile` |
|
|
|
<details> |
|
<summary> Click to expand </summary> |
|
|
|
```python |
|
import torch |
|
from transformers import AutoTokenizer, AutoModelForCausalLM |
|
|
|
tokenizer = AutoTokenizer.from_pretrained("tiiuae/Falcon3-7B-Base") |
|
model = AutoModelForCausalLM.from_pretrained("tiiuae/Falcon3-7B-Base", torch_dtype=torch.bfloat16).to(0) |
|
|
|
model = torch.compile(model) |
|
|
|
input_text = "Question: How many hours in one day? Answer: " |
|
input_ids = tokenizer(input_text, return_tensors="pt").input_ids.to("cuda") |
|
|
|
outputs = model.generate(input_ids) |
|
print(tokenizer.decode(outputs[0])) |
|
``` |
|
|
|
</details> |
|
|
|
|
|
# Training Details |
|
|
|
## Training Data |
|
|
|
Falcon3-7B is trained on 15 Gigatokens of datasets comprising of web, code, STEM, high quality and mutlilingual data. |
|
|
|
## Training Procedure |
|
|
|
Falcon3-7B is trained on 256 H100 nodes (world size 2048). |
|
|
|
### Training Hyperparameters |
|
|
|
| **Hyperparameter** | **Value** | **Comment** | |
|
|--------------------|------------|---------------------------------------| |
|
| Precision | `bfloat16` | | |
|
| Optimizer | AdamW | | |
|
| Max learning rate | 6e-4 | Following a WSD (warmup-stable-decay) | |
|
| | | learning rate scheduler | |
|
| Weight decay | 1e-1 | | |
|
| z-loss | 1e-4 | | |
|
| Batch size | Variable | Batch size was gradually increased | |
|
| | | during the training | |
|
|
|
# Evaluation |
|
|
|
<table border="1" style="width: 100%; text-align: center; border-collapse: collapse;"> |
|
<colgroup> |
|
<col style="width: 10%;"> |
|
<col style="width: 10%;"> |
|
<col style="width: 7%;"> |
|
<col style="width: 7%;"> |
|
<col style="width: 7%;"> |
|
<col style="background-color: rgba(80, 15, 213, 0.5); width: 7%;"> |
|
</colgroup> |
|
<thead> |
|
<tr> |
|
<th>Category</th> |
|
<th>Benchmark</th> |
|
<th>Llama3.1-8B</th> |
|
<th>Qwen2-7B</th> |
|
<th>Qwen2.5-7B</th> |
|
<th>Falcon3-7B-Base</th> |
|
</tr> |
|
</thead> |
|
<tbody> |
|
<tr> |
|
<td rowspan="3">General</td> |
|
<td>MMLU (5-shot)</td> |
|
<td>65.2</td> |
|
<td>70.4</td> |
|
<td>74.2</td> |
|
<td>67.5</td> |
|
</tr> |
|
<tr> |
|
<td>MMLU-PRO (5-shot)</td> |
|
<td>32.7</td> |
|
<td>42.1</td> |
|
<td>43.5</td> |
|
<td>39.2</td> |
|
</tr> |
|
<tr> |
|
<td>IFEval</td> |
|
<td>12.0</td> |
|
<td>30.6</td> |
|
<td>33.9</td> |
|
<td>34.3</td> |
|
</tr> |
|
<tr> |
|
<td rowspan="2">Math</td> |
|
<td>GSM8K (5-shot)</td> |
|
<td>49.4</td> |
|
<td>77.9</td> |
|
<td>82.9</td> |
|
<td>76.2</td> |
|
</tr> |
|
<tr> |
|
<td>MATH(4-shot)</td> |
|
<td>4.1</td> |
|
<td>17.5</td> |
|
<td>15.5</td> |
|
<td>18.0</td> |
|
</tr> |
|
<tr> |
|
<td rowspan="4">Reasoning</td> |
|
<td>Arc Challenge (25-shot)</td> |
|
<td>53.4</td> |
|
<td>57.4</td> |
|
<td>59.0</td> |
|
<td>59.6</td> |
|
</tr> |
|
<tr> |
|
<td>GPQA (0-shot)</td> |
|
<td>31.0</td> |
|
<td>31.9</td> |
|
<td>33.0</td> |
|
<td>35.5</td> |
|
</tr> |
|
<tr> |
|
<td>MUSR (0-shot)</td> |
|
<td>38.0</td> |
|
<td>44.1</td> |
|
<td>44.2</td> |
|
<td>47.3</td> |
|
</tr> |
|
<tr> |
|
<td>BBH (3-shot)</td> |
|
<td>46.5</td> |
|
<td>53.3</td> |
|
<td>54.0</td> |
|
<td>51.0</td> |
|
</tr> |
|
<tr> |
|
<td rowspan="4">CommonSense Understanding</td> |
|
<td>PIQA (0-shot)</td> |
|
<td>80.3</td> |
|
<td>79.8</td> |
|
<td>78.7</td> |
|
<td>77.7</td> |
|
</tr> |
|
<tr> |
|
<td>SciQ (0-shot)</td> |
|
<td>96.3</td> |
|
<td>95.9</td> |
|
<td>96.6</td> |
|
<td>95.3</td> |
|
</tr> |
|
<tr> |
|
<td>Winogrande (0-shot)</td> |
|
<td>74.0</td> |
|
<td>72.1</td> |
|
<td>72.9</td> |
|
<td>71.0</td> |
|
</tr> |
|
<tr> |
|
<td>OpenbookQA (0-shot)</td> |
|
<td>33.4</td> |
|
<td>35.2</td> |
|
<td>33.6</td> |
|
<td>31.4</td> |
|
</tr> |
|
</tbody> |
|
</table> |
|
|
|
|
|
|
|
# Citation |
|
|
|
|
|
|
