base_model: meta-llama/Meta-Llama-3-8B
model_type: LlamaForCausalLM
tokenizer_type: AutoTokenizer

load_in_8bit: false
load_in_4bit: false
strict: false

chat_template: chatml
datasets:
  - path: data/merged_all.json
    ds_type: json
    type: alpaca
    conversation: chatml

  - path: data/gpteacher-instruct-special-alpaca.json
    ds_type: json
    type: gpteacher
    conversation: chatml

  - path: data/wizardlm_evol_instruct_70k_random_half.json
    ds_type: json
    type: alpaca
    conversation: chatml

  - path: data/capybara_sharegpt.json
    ds_type: json
    type: sharegpt
    conversation: chatml

  - path: data/synthia-v1.3_sharegpt_12500.json
    ds_type: json
    type: sharegpt
    conversation: chatml  

  - path: data/cot_alpaca_gpt4_extracted_openhermes_2.5_sharegpt.json
    ds_type: json
    type: sharegpt
    conversation: chatml

  - path: data/slimorca_dedup_filtered_95k_sharegpt.json
    ds_type: json
    type: sharegpt
    conversation: chatml  

  - path: data/airoboros_3.2_without_contextual_slimorca_orca_sharegpt.json
    ds_type: json
    type: sharegpt
    conversation: chatml  

  - path: data/allenai_wild_chat_gpt4_english_toxic_random_half_4k_sharegpt.json
    ds_type: json
    type: sharegpt
    strict: false
    conversation: chatml  

  - path: data/pippa_bagel_repo_3k_sharegpt.json
    ds_type: json
    type: sharegpt
    conversation: chatml  

  - path: data/gpt4_data_lmys_1m_sharegpt.json
    ds_type: json
    type: sharegpt
    conversation: chatml  

  - path: data/sharegpt_gpt4_english.json
    ds_type: json
    type: sharegpt
    conversation: chatml

  - path: data/no_robots_sharegpt.json
    ds_type: json
    type: sharegpt
    strict: false
    conversation: chatml

  - path: data/oasst_top1_from_fusechatmixture_sharegpt.json
    ds_type: json
    type: sharegpt
    strict: false
    conversation: chatml

  - path: data/everythinglm-data-v3_sharegpt.json
    ds_type: json
    type: sharegpt
    strict: false
    conversation: chatml

dataset_prepared_path: last_run_prepared
val_set_size: 0.002

output_dir: ./Einstein-v6.1-Llama3-8B-model

sequence_len: 8192
sample_packing: true
pad_to_sequence_len: true
eval_sample_packing: false

wandb_project: Einstein
wandb_entity:
wandb_watch:
wandb_name: Einstein-v6.1-Llama3-2-epoch
wandb_log_model:
hub_model_id: Weyaxi/Einstein-v6.1-Llama3-8B

save_safetensors: true

gradient_accumulation_steps: 4
micro_batch_size: 1
num_epochs: 2
optimizer: adamw_bnb_8bit # look
lr_scheduler: cosine
learning_rate: 0.000005 # look

train_on_inputs: false
group_by_length: false
bf16: true
fp16: false
tf32: false

gradient_checkpointing: true
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true

warmup_steps: 10
evals_per_epoch: 2
eval_table_size:
eval_table_max_new_tokens: 128
saves_per_epoch: 2
debug:

deepspeed: zero3_bf16_cpuoffload_params.json
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:
  bos_token: "<s>"
  eos_token: "<|im_end|>"
  unk_token: "<unk>"
  pad_token: <|end_of_text|> # changed
tokens:
  - "<|im_start|>"

💬 Prompt Template

You can use ChatML prompt template while using the model:

ChatML

<|im_start|>system
{system}<|im_end|>
<|im_start|>user
{user}<|im_end|>
<|im_start|>assistant
{asistant}<|im_end|>

This prompt template is available as a chat template, which means you can format messages using the tokenizer.apply_chat_template() method:

python

messages = [
    {"role": "system", "content": "You are helpful AI asistant."},
    {"role": "user", "content": "Hello!"}
]
gen_input = tokenizer.apply_chat_template(message, return_tensors="pt")
model.generate(**gen_input)

📊 Datasets used in this model

The datasets used to train this model are listed in the metadata section of the model card.

Please note that certain datasets mentioned in the metadata may have undergone filtering based on various criteria.

The results of this filtering process and its outcomes are in the data folder of this repository:

Weyaxi/Einstein-v6.1-Llama3-8B/data

🔄 Quantizationed versions

GGUF @bartowski

https://huggingface.co/bartowski/Einstein-v6.1-Llama3-8B-GGUF

ExLlamaV2 @bartowski

https://huggingface.co/bartowski/Einstein-v6.1-Llama3-8B-exl2

AWQ @solidrust

https://huggingface.co/solidrust/Einstein-v6.1-Llama3-8B-AWQ

🎯 Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	68.60
AI2 Reasoning Challenge (25-Shot)	62.46
HellaSwag (10-Shot)	82.41
MMLU (5-Shot)	66.19
TruthfulQA (0-shot)	55.10
Winogrande (5-shot)	79.32
GSM8k (5-shot)	66.11

🎯 Open LLM Leaderboard v2 Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	19.99
IFEval (0-Shot)	45.68
BBH (3-Shot)	29.38
MATH Lvl 5 (4-Shot)	5.74
GPQA (0-shot)	4.25
MuSR (0-shot)	11.23
MMLU-PRO (5-shot)	23.68

📚 Some resources, discussions and reviews aboout this model

📱 Octopus-V4-3B

Octopus-V4-3B leverages the incredible physics capabilities of Einstein-v6.1-Llama3-8B in their model.

🤖 Additional information about training

This model is full fine-tuned for 2 epoch.

Total number of steps was 2026.

Loss graph

🤝 Acknowledgments

Thanks to sablo.ai for sponsoring this model.

Thanks to all the dataset authors mentioned in the datasets section.

Thanks to axolotl for making the repository I used to make this model.

Thanks to all open source AI community.

If you would like to support me:

☕ Buy Me a Coffee

Weyaxi/Einstein-v6.1-Llama3-8B

🔬 Einstein-v6.1-Llama3-8B