Instructions to use ea-dev/eval-agent_qwen2.5-3b_instruct_ckpt100 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ea-dev/eval-agent_qwen2.5-3b_instruct_ckpt100 with Transformers:

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

pipe = pipeline("text-generation", model="ea-dev/eval-agent_qwen2.5-3b_instruct_ckpt100")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("ea-dev/eval-agent_qwen2.5-3b_instruct_ckpt100")
model = AutoModelForCausalLM.from_pretrained("ea-dev/eval-agent_qwen2.5-3b_instruct_ckpt100", 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 ea-dev/eval-agent_qwen2.5-3b_instruct_ckpt100 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ea-dev/eval-agent_qwen2.5-3b_instruct_ckpt100"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ea-dev/eval-agent_qwen2.5-3b_instruct_ckpt100",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ea-dev/eval-agent_qwen2.5-3b_instruct_ckpt100

SGLang

How to use ea-dev/eval-agent_qwen2.5-3b_instruct_ckpt100 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 "ea-dev/eval-agent_qwen2.5-3b_instruct_ckpt100" \
    --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": "ea-dev/eval-agent_qwen2.5-3b_instruct_ckpt100",
		"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 "ea-dev/eval-agent_qwen2.5-3b_instruct_ckpt100" \
        --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": "ea-dev/eval-agent_qwen2.5-3b_instruct_ckpt100",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use ea-dev/eval-agent_qwen2.5-3b_instruct_ckpt100 with Docker Model Runner:
```
docker model run hf.co/ea-dev/eval-agent_qwen2.5-3b_instruct_ckpt100
```

eval-agent_qwen2.5-3b_instruct_ckpt100 / trainer_state.json

shulin16

Upload checkpoint 100 from eval-agent training

36737c6 verified 12 months ago

Raw

History Blame Contribute Delete

2.5 kB

	{
	"best_global_step": null,
	"best_metric": null,
	"best_model_checkpoint": null,
	"epoch": 0.6389776357827476,
	"eval_steps": 500,
	"global_step": 100,
	"is_hyper_param_search": false,
	"is_local_process_zero": true,
	"is_world_process_zero": true,
	"log_history": [
	{
	"epoch": 0.06389776357827476,
	"grad_norm": 34.803749574601504,
	"learning_rate": 1.8750000000000003e-06,
	"loss": 3.2079,
	"step": 10
	},
	{
	"epoch": 0.12779552715654952,
	"grad_norm": 8.393596619540387,
	"learning_rate": 3.958333333333333e-06,
	"loss": 2.3077,
	"step": 20
	},
	{
	"epoch": 0.19169329073482427,
	"grad_norm": 4.54582398918815,
	"learning_rate": 6.041666666666667e-06,
	"loss": 1.6795,
	"step": 30
	},
	{
	"epoch": 0.25559105431309903,
	"grad_norm": 2.88137111898239,
	"learning_rate": 8.125000000000001e-06,
	"loss": 1.3798,
	"step": 40
	},
	{
	"epoch": 0.3194888178913738,
	"grad_norm": 3.470416724559835,
	"learning_rate": 9.999862102299874e-06,
	"loss": 1.2363,
	"step": 50
	},
	{
	"epoch": 0.38338658146964855,
	"grad_norm": 3.111739768988857,
	"learning_rate": 9.983323579940351e-06,
	"loss": 1.1359,
	"step": 60
	},
	{
	"epoch": 0.4472843450479233,
	"grad_norm": 2.63691114437442,
	"learning_rate": 9.939310009499348e-06,
	"loss": 1.0504,
	"step": 70
	},
	{
	"epoch": 0.5111821086261981,
	"grad_norm": 2.3175377091892346,
	"learning_rate": 9.868064055324204e-06,
	"loss": 1.0116,
	"step": 80
	},
	{
	"epoch": 0.5750798722044729,
	"grad_norm": 2.30928876775904,
	"learning_rate": 9.76997852474223e-06,
	"loss": 0.9679,
	"step": 90
	},
	{
	"epoch": 0.6389776357827476,
	"grad_norm": 2.2162135977370196,
	"learning_rate": 9.645594202357438e-06,
	"loss": 0.944,
	"step": 100
	}
	],
	"logging_steps": 10,
	"max_steps": 471,
	"num_input_tokens_seen": 0,
	"num_train_epochs": 3,
	"save_steps": 100,
	"stateful_callbacks": {
	"TrainerControl": {
	"args": {
	"should_epoch_stop": false,
	"should_evaluate": false,
	"should_log": false,
	"should_save": true,
	"should_training_stop": false
	},
	"attributes": {}
	}
	},
	"total_flos": 7758272790528.0,
	"train_batch_size": 4,
	"trial_name": null,
	"trial_params": null
	}