Instructions to use mfirth/agi-ds with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use mfirth/agi-ds with Transformers:

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

pipe = pipeline("text-generation", model="mfirth/agi-ds", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("mfirth/agi-ds", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("mfirth/agi-ds", trust_remote_code=True)
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 Settings

vLLM

How to use mfirth/agi-ds with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/mfirth/agi-ds

SGLang

How to use mfirth/agi-ds 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 "mfirth/agi-ds" \
    --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": "mfirth/agi-ds",
		"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 "mfirth/agi-ds" \
        --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": "mfirth/agi-ds",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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

See axolotl config

axolotl version: 0.5.3.dev44+g5bef1906

base_model: deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct
trust_remote_code: true

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

datasets:
  - path: axolotl_format_deepseek_combined_wm.json
    type: input_output
dataset_prepared_path: last_run_prepared_deepseek
    
output_dir: ./models/deepseek_wm
sequence_len: 4096

wandb_project: agent-v0
wandb_name: deepseek_wm

train_on_inputs: false
gradient_checkpointing: true
gradient_checkpointing_kwargs:
  use_reentrant: false
gradient_accumulation_steps: 1
micro_batch_size: 4
num_epochs: 3
optimizer: adamw_torch
learning_rate: 2e-5
xformers_attention:
flash_attention: true

logging_steps: 5

warmup_steps: 5
saves_per_epoch: 1
weight_decay: 0.0

deepspeed: axolotl/deepspeed_configs/zero3_bf16_cpuoffload_all.json

models/deepseek_wm

This model is a fine-tuned version of deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct on the axolotl_format_deepseek_combined_wm.json dataset.

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: 2e-05
train_batch_size: 4
eval_batch_size: 4
seed: 42
distributed_type: multi-GPU
num_devices: 8
total_train_batch_size: 32
total_eval_batch_size: 32
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 5
num_epochs: 3

Training results

Framework versions

Transformers 4.47.0
Pytorch 2.5.1+cu124
Datasets 3.1.0
Tokenizers 0.21.0

Downloads last month: 5

Safetensors

Model size

16B params

Tensor type

BF16

Model tree for mfirth/agi-ds

Base model

deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct

Finetuned

(17)

this model

Quantizations

1 model