--- license: other license_name: llama-3 license_link: https://huggingface.co./meta-llama/Meta-Llama-3-8B-Instruct/raw/main/LICENSE base_model: meta-llama/Meta-Llama-3-8B-Instruct tags: - generated_from_trainer model-index: - name: lightblue/suzume-llama-3-8B-multilingual results: [] ---

Suzume - a Japanese tree sparrow

# Suzume [[Paper](https://arxiv.org/abs/2405.12612)] [[Dataset](https://huggingface.co./datasets/lightblue/tagengo-gpt4)] This Suzume 8B, a multilingual finetune of Llama 3. Llama 3 has exhibited excellent performance on many English language benchmarks. However, it also seemingly been finetuned on mostly English data, meaning that it will respond in English, even if prompted in other languages. We have fine-tuned Llama 3 on more than 80,000 multilingual conversations meaning that this model has the smarts of Llama 3 but has the added ability to chat in more languages. Please feel free to comment on this model and give us feedback in the Community tab! # How to use The easiest way to use this model on your own computer is to use the [GGUF version of this model (lightblue/suzume-llama-3-8B-multilingual-gguf)](https://huggingface.co./lightblue/suzume-llama-3-8B-multilingual-gguf) using a program such as [jan.ai](https://jan.ai/) or [LM Studio](https://lmstudio.ai/). If you want to use this model directly in Python, we recommend using vLLM for the fastest inference speeds. ```python from vllm import LLM, SamplingParams sampling_params = SamplingParams(temperature=0.0, max_tokens=100) llm = LLM(model="lightblue/suzume-llama-3-8B-multilingual") messages = [] messages.append({"role": "user", "content": "Bonjour!"}) prompt = llm.llm_engine.tokenizer.tokenizer.apply_chat_template(conversation=messages, add_generation_prompt=True, tokenize=False) prompts = [prompt] outputs = llm.generate(prompts, sampling_params) for output in outputs: prompt = output.prompt generated_text = output.outputs[0].text print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}") ``` # Evaluation scores We achieve the following MT-Bench scores across 6 languages: | | **meta-llama/Meta-Llama-3-8B-Instruct** | **lightblue/suzume-llama-3-8B-multilingual** | **Nexusflow/Starling-LM-7B-beta** | **gpt-3.5-turbo** | |-----------------|-----------------------------------------|----------------------------------------------|-----------------------------------|-------------------| | **German** πŸ‡©πŸ‡ͺ | NaN | 7.26 | 6.99 | 7.68 | | **French** πŸ‡«πŸ‡· | NaN | 7.66 | 7.29 | 7.74 | | **Japanese** πŸ‡―πŸ‡΅ | NaN | 6.56 | 6.22 | 7.84 | | **Russian** πŸ‡·πŸ‡Ί | NaN | 8.19 | 8.28 | 7.94 | | **Chinese** πŸ‡¨πŸ‡³ | NaN | 7.11 | 6.97 | 7.55 | | **English** πŸ‡ΊπŸ‡Έ | 7.98 | 7.73 | 7.92 | 8.26 | We observe minimal degredation of Llama 3's English ability while achieving best-in-class multilingual abilities compared to the top rated 7B model ([Nexusflow/Starling-LM-7B-beta](https://huggingface.co./Nexusflow/Starling-LM-7B-beta)) on the [Chatbot Arena Leaderboard](https://chat.lmsys.org/?leaderboard). [Here is our evaluation script.](https://drive.google.com/file/d/15HPn7452t8LbTD9HKSl7ngYYWnsoOG08/view?usp=sharing) # Training data We train on three sources of data to create this model: * [lightblue/tagengo-gpt4](https://huggingface.co./datasets/lightblue/tagengo-gpt4) - 76,338 conversations * A diverse dataset of initial inputs sampled from [lmsys/lmsys-chat-1m](https://huggingface.co./datasets/lmsys/lmsys-chat-1m) and then used to prompt `gpt-4-0125-preview` * [megagonlabs/instruction_ja](https://github.com/megagonlabs/instruction_ja) - 669 conversations * A hand-edited dataset of nearly 700 Japanese conversations taken originally from translations of the [kunishou/hh-rlhf-49k-ja](https://huggingface.co./datasets/kunishou/hh-rlhf-49k-ja) dataset. * [openchat/openchat_sharegpt4_dataset](https://huggingface.co./datasets/openchat/openchat_sharegpt4_dataset/resolve/main/sharegpt_gpt4.json) - 6,206 conversations * Multilingual conversations of humans talking to GPT-4.
We prepare our data like so: ```python import pandas as pd from datasets import Dataset, load_dataset, concatenate_datasets ### Tagengo gpt4_dataset = load_dataset("lightblue/tagengo-gpt4", split="train") gpt4_dataset = gpt4_dataset.filter(lambda x: x["response"][1] == "stop") #### ### Megagon megagon_df = pd.read_json( "https://raw.githubusercontent.com/megagonlabs/instruction_ja/main/data/data.jsonl", lines=True, orient="records" ) role_map = {"user": "human", "agent": "gpt"} megagon_df["conversations"] = megagon_df.utterances.apply(lambda x: [{"from": role_map[y["name"]], "value": y["text"]} for y in x]) megagon_df["language"] = "Japanese" megagon_df = megagon_df[["conversations", "language"]] megagon_dataset = Dataset.from_pandas(df) ### ### Openchat openchat_df = pd.read_json("https://huggingface.co./datasets/openchat/openchat_sharegpt4_dataset/resolve/main/sharegpt_gpt4.json?download=true") openchat_df["conversations"] = openchat_df["items"] openchat_dataset = Dataset.from_pandas(openchat_df) ### dataset = concatenate_datasets([gpt4_dataset, megagon_dataset, openchat_dataset]) dataset = dataset.filter(lambda x: not any([y["value"] is None for y in x["conversations"]])) dataset.select_columns(["conversations"]).to_json("/workspace/llm_training/axolotl/llama3-multilingual/tagengo_openchat_megagon.json") ```

# workspace/llm_training/axolotl/llama3-multilingual/output_tagengo_openchat_megagon_8B_llama3 This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co./meta-llama/Meta-Llama-3-8B-Instruct) on the above described dataset. It achieves the following results on the evaluation set: - Loss: 0.6595 ## Training procedure [Built with Axolotl](https://github.com/OpenAccess-AI-Collective/axolotl)
See axolotl config axolotl version: `0.4.0` ```yaml base_model: meta-llama/Meta-Llama-3-8B-Instruct model_type: LlamaForCausalLM tokenizer_type: AutoTokenizer # PreTrainedTokenizerFast load_in_8bit: false load_in_4bit: false strict: false datasets: - path: /workspace/llm_training/axolotl/llama3-multilingual/tagengo_openchat_megagon.json ds_type: json # see other options below type: sharegpt conversation: llama-3 dataset_prepared_path: /workspace/llm_training/axolotl/llama3-multilingual/prepared_tagengo_openchat_megagon val_set_size: 0.01 output_dir: /workspace/llm_training/axolotl/llama3-multilingual/output_tagengo_openchat_megagon_8B_llama3 sequence_len: 8192 sample_packing: true pad_to_sequence_len: true use_wandb: true wandb_project: wandb_project wandb_entity: wandb_entity wandb_name: wandb_name gradient_accumulation_steps: 2 micro_batch_size: 2 num_epochs: 1 optimizer: paged_adamw_8bit lr_scheduler: cosine learning_rate: 1e-5 train_on_inputs: false group_by_length: false bf16: auto fp16: tf32: false gradient_checkpointing: true gradient_checkpointing_kwargs: use_reentrant: false early_stopping_patience: resume_from_checkpoint: logging_steps: 1 xformers_attention: flash_attention: true warmup_steps: 10 evals_per_epoch: 5 eval_table_size: saves_per_epoch: 1 debug: deepspeed: /workspace/axolotl/deepspeed_configs/zero2.json weight_decay: 0.0 special_tokens: pad_token: <|end_of_text|> ```

Note - we added this Llama 3 template to fastchat directly as the Llama 3 chat template was not supported when we trained this model. ```python from fastchat.conversation import Conversation from fastchat.conversation import register_conv_template from fastchat.conversation import SeparatorStyle register_conv_template( Conversation( name="llama-3", system_template="<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\n{system_message}", roles=("<|start_header_id|>user<|end_header_id|>\n", "<|start_header_id|>assistant<|end_header_id|>\n"), sep_style=SeparatorStyle.ADD_NEW_LINE_SINGLE, sep="<|eot_id|>", stop_token_ids=[128009], stop_str="<|eot_id|>", ) ) ```

### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 1e-05 - train_batch_size: 2 - eval_batch_size: 2 - seed: 42 - distributed_type: multi-GPU - num_devices: 4 - gradient_accumulation_steps: 2 - total_train_batch_size: 16 - total_eval_batch_size: 8 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: cosine - lr_scheduler_warmup_steps: 10 - num_epochs: 1 ### Training results | Training Loss | Epoch | Step | Validation Loss | |:-------------:|:-----:|:----:|:---------------:| | 1.1894 | 0.0 | 1 | 1.0110 | | 0.8493 | 0.2 | 73 | 0.7057 | | 0.8047 | 0.4 | 146 | 0.6835 | | 0.7644 | 0.6 | 219 | 0.6687 | | 0.7528 | 0.8 | 292 | 0.6615 | | 0.7794 | 1.0 | 365 | 0.6595 | ### Framework versions - Transformers 4.38.2 - Pytorch 2.2.1+cu121 - Datasets 2.18.0 - Tokenizers 0.15.0 # How to cite Please cite [this paper](https://arxiv.org/abs/2405.12612) when referencing this model. ```tex @article{devine2024tagengo, title={Tagengo: A Multilingual Chat Dataset}, author={Devine, Peter}, journal={arXiv preprint arXiv:2405.12612}, year={2024} } ``` # Developer Peter Devine - ([ptrdvn](https://huggingface.co./ptrdvn))