You can deploy this model and start earning money today!

RichardErkhov/
flammenai_-_Mahou-1.2a-llama3-8B-gguf

Name: flammenai_-_Mahou-1.2a-llama3-8B-gguf
Availability: InStock

Views

No views yet

Quantization made by Richard Erkhov.

Github

Discord

Request more models

Mahou-1.2a-llama3-8B - GGUF

Model creator: https://huggingface.co/flammenai/
Original model: https://huggingface.co/flammenai/Mahou-1.2a-llama3-8B/

Name	Quant method	Size
Mahou-1.2a-llama3-8B.Q2_K.gguf	Q2_K	2.96GB
Mahou-1.2a-llama3-8B.IQ3_XS.gguf	IQ3_XS	3.28GB
Mahou-1.2a-llama3-8B.IQ3_S.gguf	IQ3_S	3.43GB
Mahou-1.2a-llama3-8B.Q3_K_S.gguf	Q3_K_S	3.41GB
Mahou-1.2a-llama3-8B.IQ3_M.gguf	IQ3_M	3.52GB
Mahou-1.2a-llama3-8B.Q3_K.gguf	Q3_K	3.74GB
Mahou-1.2a-llama3-8B.Q3_K_M.gguf	Q3_K_M	3.74GB
Mahou-1.2a-llama3-8B.Q3_K_L.gguf	Q3_K_L	4.03GB
Mahou-1.2a-llama3-8B.IQ4_XS.gguf	IQ4_XS	4.18GB
Mahou-1.2a-llama3-8B.Q4_0.gguf	Q4_0	4.34GB
Mahou-1.2a-llama3-8B.IQ4_NL.gguf	IQ4_NL	4.38GB
Mahou-1.2a-llama3-8B.Q4_K_S.gguf	Q4_K_S	4.37GB
Mahou-1.2a-llama3-8B.Q4_K.gguf	Q4_K	4.58GB
Mahou-1.2a-llama3-8B.Q4_K_M.gguf	Q4_K_M	4.58GB
Mahou-1.2a-llama3-8B.Q4_1.gguf	Q4_1	4.78GB
Mahou-1.2a-llama3-8B.Q5_0.gguf	Q5_0	5.21GB
Mahou-1.2a-llama3-8B.Q5_K_S.gguf	Q5_K_S	5.21GB
Mahou-1.2a-llama3-8B.Q5_K.gguf	Q5_K	5.34GB
Mahou-1.2a-llama3-8B.Q5_K_M.gguf	Q5_K_M	5.34GB
Mahou-1.2a-llama3-8B.Q5_1.gguf	Q5_1	5.65GB
Mahou-1.2a-llama3-8B.Q6_K.gguf	Q6_K	6.14GB
Mahou-1.2a-llama3-8B.Q8_0.gguf	Q8_0	7.95GB

Original model description:

license: llama3 library_name: transformers base_model:

nbeerbower/llama-3-Stheno-Mahou-8B datasets:
flammenai/FlameMix-DPO-v1
flammenai/Grill-preprod-v1_chatML
flammenai/Grill-preprod-v2_chatML model-index:
name: Mahou-1.2a-llama3-8B results:
- task: type: text-generation name: Text Generation dataset: name: IFEval (0-Shot) type: HuggingFaceH4/ifeval args: num_few_shot: 0 metrics:
  - type: inst_level_strict_acc and prompt_level_strict_acc value: 50.93 name: strict accuracy source: url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=flammenai/Mahou-1.2a-llama3-8B name: Open LLM Leaderboard
- task: type: text-generation name: Text Generation dataset: name: BBH (3-Shot) type: BBH args: num_few_shot: 3 metrics:
  - type: acc_norm value: 28.97 name: normalized accuracy source: url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=flammenai/Mahou-1.2a-llama3-8B name: Open LLM Leaderboard
- task: type: text-generation name: Text Generation dataset: name: MATH Lvl 5 (4-Shot) type: hendrycks/competition_math args: num_few_shot: 4 metrics:
  - type: exact_match value: 7.55 name: exact match source: url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=flammenai/Mahou-1.2a-llama3-8B name: Open LLM Leaderboard
- task: type: text-generation name: Text Generation dataset: name: GPQA (0-shot) type: Idavidrein/gpqa args: num_few_shot: 0 metrics:
  - type: acc_norm value: 5.15 name: acc_norm source: url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=flammenai/Mahou-1.2a-llama3-8B name: Open LLM Leaderboard
- task: type: text-generation name: Text Generation dataset: name: MuSR (0-shot) type: TAUR-Lab/MuSR args: num_few_shot: 0 metrics:
  - type: acc_norm value: 6.02 name: acc_norm source: url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=flammenai/Mahou-1.2a-llama3-8B name: Open LLM Leaderboard
- task: type: text-generation name: Text Generation dataset: name: MMLU-PRO (5-shot) type: TIGER-Lab/MMLU-Pro config: main split: test args: num_few_shot: 5 metrics:
  - type: acc value: 31.3 name: accuracy source: url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=flammenai/Mahou-1.2a-llama3-8B name: Open LLM Leaderboard

Mahou-1.2a-llama3-8B

Mahou is our attempt to build a production-ready conversational/roleplay LLM.

Future versions will be released iteratively and finetuned from flammen.ai conversational data.

Chat Format

This model has been trained to use ChatML format.

<|im_start|>system
{{system}}<|im_end|>
<|im_start|>{{char}}
{{message}}<|im_end|>
<|im_start|>{{user}}
{{message}}<|im_end|>

Roleplay Format

Speech without quotes.
Actions in *asterisks*

*leans against wall cooly* so like, i just casted a super strong spell at magician academy today, not gonna lie, felt badass.

ST Settings

Use ChatML for the Context Template.
Turn on Instruct Mode for ChatML.
Use the following stopping strings: ["<", "|", "<|", "\n"]

Method

Finetuned using an A100 on Google Colab.

Fine-tune a Mistral-7b model with Direct Preference Optimization - Maxime Labonne

Configuration

LoRA, model, and training settings:

python

# LoRA configuration
peft_config = LoraConfig(
    r=16,
    lora_alpha=16,
    lora_dropout=0.05,
    bias="none",
    task_type="CAUSAL_LM",
    target_modules=['k_proj', 'gate_proj', 'v_proj', 'up_proj', 'q_proj', 'o_proj', 'down_proj']
)

# Model to fine-tune
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype=torch.bfloat16,
    load_in_4bit=True
)
model.config.use_cache = False

# Reference model
ref_model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype=torch.bfloat16,
    load_in_4bit=True
)

# Training arguments
training_args = TrainingArguments(
    per_device_train_batch_size=4,
    gradient_accumulation_steps=4,
    gradient_checkpointing=True,
    learning_rate=5e-5,
    lr_scheduler_type="cosine",
    max_steps=2000,
    save_strategy="no",
    logging_steps=1,
    output_dir=new_model,
    optim="paged_adamw_32bit",
    warmup_steps=100,
    bf16=True,
    report_to="wandb",
)

# Create DPO trainer
dpo_trainer = DPOTrainer(
    model,
    ref_model,
    args=training_args,
    train_dataset=dataset,
    tokenizer=tokenizer,
    peft_config=peft_config,
    beta=0.1,
    force_use_ref_model=True
)

# Fine-tune model with DPO
dpo_trainer.train()

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	21.65
IFEval (0-Shot)	50.93
BBH (3-Shot)	28.97
MATH Lvl 5 (4-Shot)	7.55
GPQA (0-shot)	5.15
MuSR (0-shot)	6.02
MMLU-PRO (5-shot)	31.30

RichardErkhov/flammenai_-_Mahou-1.2a-llama3-8B-gguf