Instructions to use FourOhFour/Fatgirl_v2_8B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use FourOhFour/Fatgirl_v2_8B with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="FourOhFour/Fatgirl_v2_8B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("FourOhFour/Fatgirl_v2_8B")
model = AutoModelForCausalLM.from_pretrained("FourOhFour/Fatgirl_v2_8B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use FourOhFour/Fatgirl_v2_8B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "FourOhFour/Fatgirl_v2_8B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "FourOhFour/Fatgirl_v2_8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/FourOhFour/Fatgirl_v2_8B

SGLang

How to use FourOhFour/Fatgirl_v2_8B with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "FourOhFour/Fatgirl_v2_8B" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "FourOhFour/Fatgirl_v2_8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "FourOhFour/Fatgirl_v2_8B" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "FourOhFour/Fatgirl_v2_8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use FourOhFour/Fatgirl_v2_8B with Docker Model Runner:
```
docker model run hf.co/FourOhFour/Fatgirl_v2_8B
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

See axolotl config

axolotl version: 0.4.1

base_model: FourOhFour/Magic_v2_8B
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer

load_in_8bit: false
load_in_4bit: false
strict: false

datasets:
  - path: Epiculous/SynthRP-Gens-v1.1-Filtered-n-Cleaned
    type: sharegpt
    conversation: chatml
  - path: ResplendentAI/bluemoon
    type: sharegpt
    conversation: chatml
  - path: openerotica/freedom-rp
    type: sharegpt
    conversation: chatml
  - path: MinervaAI/Aesir-Preview
    type: sharegpt
    conversation: chatml
  - path: anthracite-core/c2_logs_32k_v1.1
    type: sharegpt
    conversation: chatml
  - path: Nitral-AI/Creative_Writing-ShareGPT
    type: sharegpt
    conversation: chatml
  - path: PJMixers/lodrick-the-lafted_OpusStories-Story2Prompt-ShareGPT
    type: sharegpt
    conversation: chatml

chat_template: chatml

val_set_size: 0.002
output_dir: ./outputs/out

adapter:
lora_r:
lora_alpha:
lora_dropout:
lora_target_linear:

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

plugins:
  - axolotl.integrations.liger.LigerPlugin
liger_rope: true
liger_rms_norm: true
liger_swiglu: true
liger_fused_linear_cross_entropy: true

wandb_project: mini8B
wandb_entity:
wandb_watch:
wandb_name: mini8B
wandb_log_model:

gradient_accumulation_steps: 8
micro_batch_size: 2
num_epochs: 2
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 0.00001
weight_decay: 0.05

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

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

warmup_ratio: 0.1
evals_per_epoch: 4
eval_table_size:
eval_max_new_tokens: 128
saves_per_epoch: 2

debug:
deepspeed: deepspeed_configs/zero3_bf16.json
fsdp:
fsdp_config:

special_tokens:
  pad_token: <pad>

outputs/out

This model is a fine-tuned version of FourOhFour/Magic_v2_8B on the None dataset. It achieves the following results on the evaluation set:

Loss: 2.6845

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

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: 2
gradient_accumulation_steps: 8
total_train_batch_size: 32
total_eval_batch_size: 4
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 58
num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss
1.7471	0.0034	1	2.8918
1.5602	0.2507	74	2.7319
1.4587	0.5015	148	2.6953
1.5022	0.7522	222	2.6729
1.4152	1.0030	296	2.6487
1.2528	1.2501	370	2.6922
1.2245	1.5002	444	2.6843
1.2803	1.7503	518	2.6845