Instructions to use PrimeIntellect/GLM-0.5B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use PrimeIntellect/GLM-0.5B with Transformers:

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

pipe = pipeline("text-generation", model="PrimeIntellect/GLM-0.5B", 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("PrimeIntellect/GLM-0.5B", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("PrimeIntellect/GLM-0.5B", 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

vLLM

How to use PrimeIntellect/GLM-0.5B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "PrimeIntellect/GLM-0.5B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "PrimeIntellect/GLM-0.5B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/PrimeIntellect/GLM-0.5B

SGLang

How to use PrimeIntellect/GLM-0.5B 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 "PrimeIntellect/GLM-0.5B" \
    --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": "PrimeIntellect/GLM-0.5B",
		"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 "PrimeIntellect/GLM-0.5B" \
        --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": "PrimeIntellect/GLM-0.5B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use PrimeIntellect/GLM-0.5B with Docker Model Runner:
```
docker model run hf.co/PrimeIntellect/GLM-0.5B
```

GLM-0.5B / config.json

mikasenghaas

Upload folder using huggingface_hub

9da3a96 verified 8 months ago

raw

history blame contribute delete

1.2 kB

	{
	"architectures": [
	"Glm4MoeForCausalLM"
	],
	"auto_map": {
	"AutoConfig": "configuration_glm4_moe.Glm4MoeConfig",
	"AutoModelForCausalLM": "modeling_glm4_moe.Glm4MoeForCausalLM",
	"AutoModel": "modeling_glm4_moe.Glm4MoeModel"
	},
	"attention_bias": true,
	"attention_dropout": 0.0,
	"pad_token_id": 151329,
	"eos_token_id": [
	151329,
	151336,
	151338
	],
	"head_dim": 64,
	"hidden_act": "silu",
	"hidden_size": 1024,
	"partial_rotary_factor": 0.5,
	"initializer_range": 0.02,
	"intermediate_size": 2048,
	"max_position_embeddings": 131072,
	"model_type": "glm4_moe",
	"moe_intermediate_size": 256,
	"norm_topk_prob": true,
	"num_attention_heads": 16,
	"n_group": 1,
	"topk_group": 1,
	"n_routed_experts": 8,
	"n_shared_experts": 1,
	"routed_scaling_factor": 1.0,
	"num_experts_per_tok": 4,
	"first_k_dense_replace": 1,
	"num_hidden_layers": 24,
	"num_key_value_heads": 4,
	"rms_norm_eps": 1e-05,
	"rope_scaling": null,
	"rope_theta": 1000000,
	"num_nextn_predict_layers": 1,
	"tie_word_embeddings": false,
	"torch_dtype": "bfloat16",
	"transformers_version": "4.54.0",
	"use_cache": true,
	"use_qk_norm": false,
	"vocab_size": 151552
	}