File size: 7,839 Bytes

---
language:
- en
- th
license: llama3
tags:
- instruct
- chat
pipeline_tag: text-generation
model-index:
- name: llama-3-typhoon-v1.5-8b-instruct
  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: 60.41
      name: normalized accuracy
    source:
      url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=scb10x/llama-3-typhoon-v1.5-8b-instruct
      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: 80.79
      name: normalized accuracy
    source:
      url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=scb10x/llama-3-typhoon-v1.5-8b-instruct
      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: 64.46
      name: accuracy
    source:
      url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=scb10x/llama-3-typhoon-v1.5-8b-instruct
      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: 53.25
    source:
      url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=scb10x/llama-3-typhoon-v1.5-8b-instruct
      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: 77.66
      name: accuracy
    source:
      url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=scb10x/llama-3-typhoon-v1.5-8b-instruct
      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: 57.16
      name: accuracy
    source:
      url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=scb10x/llama-3-typhoon-v1.5-8b-instruct
      name: Open LLM Leaderboard
---
**Llama-3-Typhoon-v1.5-8B: Thai Large Language Model (Instruct)**

**Llama-3-Typhoon-v1.5-8B-instruct** is a *instruct* Thai 🇹🇭 large language model with 8 billion parameters, and it is based on Llama3-8B.

![Typhoon 1.5 8b benchmark](https://storage.googleapis.com/typhoon-public/assets/1.5-8b-benchmark.png)

For release post, please see our [blog](https://blog.opentyphoon.ai/typhoon-1-5-release-a9364cb8e8d7). 
*To acknowledge Meta's effort in creating the foundation model and to comply with the license, we explicitly include "llama-3" in the model name.

## **Model Description**

- **Model type**: A 8B instruct decoder-only model based on Llama architecture.
- **Requirement**: transformers 4.38.0 or newer.
- **Primary Language(s)**: Thai 🇹🇭 and English 🇬🇧
- **License**: [Llama 3 Community License](https://llama.meta.com/llama3/license/)

## **Performance**

| Model | ONET | IC | TGAT | TPAT-1 | A-Level | Average (ThaiExam) | M3Exam | MMLU |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Typhoon-1.0 (Mistral) | 0.379 | 0.393 | 0.700 | 0.414 | 0.324 | 0.442 | 0.391 | 0.547 |
| Typhoon-1.5 8B (Llama3) | ***0.446*** | ***0.431*** | ***0.722*** | ***0.526*** | ***0.407*** | ***0.506*** | ***0.460*** | ***0.614*** |
| Sailor 7B | 0.372 | 0.379 | 0.678 | 0.405 | 0.396 | 0.446 | 0.411 | 0.553 |
| SeaLLM 2.0 7B | 0.327 | 0.311 | 0.656 | 0.414 | 0.321 | 0.406 | 0.354 | 0.579 |
| OpenThaiGPT 1.0.0 7B | 0.238 | 0.249 | 0.444 | 0.319 | 0.289 | 0.308 | 0.268 | 0.369 |
| SambaLingo-Thai-Chat 7B | 0.251 | 0.241 | 0.522 | 0.302 | 0.262 | 0.316 | 0.309 | 0.388 |


## Usage Example

```python
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

model_id = "scb10x/llama-3-typhoon-v1.5-8b-instruct"

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype=torch.bfloat16,
    device_map="auto",
)

messages = [
    {"role": "system", "content": "You are a helpful assistant who're always speak Thai."},
    {"role": "user", "content": "ขอสูตรไก่ย่าง"},
]

input_ids = tokenizer.apply_chat_template(
    messages,
    add_generation_prompt=True,
    return_tensors="pt"
).to(model.device)

terminators = [
    tokenizer.eos_token_id,
    tokenizer.convert_tokens_to_ids("<|eot_id|>")
]

outputs = model.generate(
    input_ids,
    max_new_tokens=512,
    eos_token_id=terminators,
    do_sample=True,
    temperature=0.4,
    top_p=0.9,
)
response = outputs[0][input_ids.shape[-1]:]
print(tokenizer.decode(response, skip_special_tokens=True))
```

## Chat Template

We use llama3 chat-template.

```python
{% set loop_messages = messages %}{% for message in loop_messages %}{% set content = '<|start_header_id|>' + message['role'] + '<|end_header_id|>\n\n'+ message['content'] | trim + '<|eot_id|>' %}{% if loop.index0 == 0 %}{% set content = bos_token + content %}{% endif %}{{ content }}{% endfor %}{% if add_generation_prompt %}{{ '<|start_header_id|>assistant<|end_header_id|>\n\n' }}{% endif %}
```

## **Intended Uses & Limitations**

This model is an instructional model. However, it’s still undergoing development. It incorporates some level of guardrails, but it still may produce answers that are inaccurate, biased, or otherwise objectionable in response to user prompts. We recommend that developers assess these risks in the context of their use case.

## **Follow us**

**https://twitter.com/opentyphoon**

## **Support**

**https://discord.gg/CqyBscMFpg**

## **SCB10X AI Team**

- Kunat Pipatanakul, Potsawee Manakul, Sittipong Sripaisarnmongkol, Pathomporn Chokchainant, Kasima Tharnpipitchai
- If you find Typhoon-8B useful for your work, please cite it using:

```
@article{pipatanakul2023typhoon,
    title={Typhoon: Thai Large Language Models}, 
    author={Kunat Pipatanakul and Phatrasek Jirabovonvisut and Potsawee Manakul and Sittipong Sripaisarnmongkol and Ruangsak Patomwong and Pathomporn Chokchainant and Kasima Tharnpipitchai},
    year={2023},
    journal={arXiv preprint arXiv:2312.13951},
    url={https://arxiv.org/abs/2312.13951}
}
```

## **Contact Us**

- General & Collaboration: **[kasima@scb10x.com](mailto:kasima@scb10x.com)**, **[pathomporn@scb10x.com](mailto:pathomporn@scb10x.com)**
- Technical: **[kunat@scb10x.com](mailto:kunat@scb10x.com)**
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_scb10x__llama-3-typhoon-v1.5-8b-instruct)

|             Metric              |Value|
|---------------------------------|----:|
|Avg.                             |65.62|
|AI2 Reasoning Challenge (25-Shot)|60.41|
|HellaSwag (10-Shot)              |80.79|
|MMLU (5-Shot)                    |64.46|
|TruthfulQA (0-shot)              |53.25|
|Winogrande (5-shot)              |77.66|
|GSM8k (5-shot)                   |57.16|