Instructions to use Kendamarron/Qwen2.5-3b-base-split2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Kendamarron/Qwen2.5-3b-base-split2 with Transformers:

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

pipe = pipeline("text-generation", model="Kendamarron/Qwen2.5-3b-base-split2")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Kendamarron/Qwen2.5-3b-base-split2")
model = AutoModelForCausalLM.from_pretrained("Kendamarron/Qwen2.5-3b-base-split2", device_map="auto")
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 Kendamarron/Qwen2.5-3b-base-split2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Kendamarron/Qwen2.5-3b-base-split2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Kendamarron/Qwen2.5-3b-base-split2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Kendamarron/Qwen2.5-3b-base-split2

SGLang

How to use Kendamarron/Qwen2.5-3b-base-split2 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 "Kendamarron/Qwen2.5-3b-base-split2" \
    --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": "Kendamarron/Qwen2.5-3b-base-split2",
		"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 "Kendamarron/Qwen2.5-3b-base-split2" \
        --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": "Kendamarron/Qwen2.5-3b-base-split2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Kendamarron/Qwen2.5-3b-base-split2 with Docker Model Runner:
```
docker model run hf.co/Kendamarron/Qwen2.5-3b-base-split2
```

See axolotl config

axolotl version: 0.11.0.dev0

# 学習のベースモデルに関する設定
base_model: Qwen/Qwen2.5-3B
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer

# 学習後のモデルのHFへのアップロードに関する設定
hub_model_id: Kendamarron/Qwen2.5-3b-base-split2
hub_strategy: "end"
push_dataset_to_hub:
hf_use_auth_token: true

# Liger Kernelの設定（学習の軽量・高速化）
plugins:
  - axolotl.integrations.liger.LigerPlugin
liger_cross_entropy: false
liger_rope: true
liger_rms_norm: true
liger_swiglu: true
liger_fused_linear_cross_entropy: true

# 量子化に関する設定
load_in_8bit: false
load_in_4bit: false

# SFTに利用するchat templateの設定
chat_template: qwen_25

# 学習データセットの前処理に関する設定
datasets:
  - path: Kendamarron/if_inst_checked-llm-jp-3.1-8x13b-instruct4-base-split2
    split: train
    type: chat_template
    field_messages: conversations
    message_field_role: role
    message_field_content: content

# データセット、モデルの出力先に関する設定
shuffle_merged_datasets: true
dataset_prepared_path: /workspace/data/sft-data
output_dir: /workspace/data/models/Qwen2.5-3B-base-split2

# valid datasetのサイズ
val_set_size: 0

# wandbに関する設定
wandb_project: axolotl
wandb_entity: kendamarron
wandb_watch:
wandb_name: qwen2.5-3b-base-split2
wandb_log_model:

# 学習に関する様々な設定
sequence_len: 4096
sample_packing: false
eval_sample_packing: false

gradient_accumulation_steps: 4
micro_batch_size: 8
num_epochs: 1
optimizer: paged_adamw_8bit
lr_scheduler: cosine
cosine_min_lr_ratio: 0.1
learning_rate: 1e-5

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

gradient_checkpointing: false
early_stopping_patience:
auto_resume_from_checkpoints: true
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true

save_strategy: steps
save_steps: 1000
save_total_limit: 2

warmup_steps: 100
debug:
deepspeed: 
weight_decay: 0.1
fsdp:
fsdp_config:

Qwen2.5-3b-base-split2

This model is a fine-tuned version of Qwen/Qwen2.5-3B on the Kendamarron/if_inst_checked-llm-jp-3.1-8x13b-instruct4-base-split2 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: 1e-05
train_batch_size: 8
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 32
optimizer: Use OptimizerNames.PAGED_ADAMW_8BIT 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: 100
training_steps: 465

Training results

Framework versions

Transformers 4.52.4
Pytorch 2.6.0+cu124
Datasets 3.6.0
Tokenizers 0.21.1

Downloads last month: 6

Safetensors

Model size

3B params

Tensor type

BF16

Model tree for Kendamarron/Qwen2.5-3b-base-split2

Base model

Qwen/Qwen2.5-3B

Finetuned

(525)

this model