Instructions to use tiny-random/glm-5.3-flash with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tiny-random/glm-5.3-flash with Transformers:

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

pipe = pipeline("image-text-to-text", model="tiny-random/glm-5.3-flash")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("tiny-random/glm-5.3-flash")
model = AutoModelForMultimodalLM.from_pretrained("tiny-random/glm-5.3-flash", device_map="auto")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.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=256)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use tiny-random/glm-5.3-flash with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "tiny-random/glm-5.3-flash"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tiny-random/glm-5.3-flash",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/tiny-random/glm-5.3-flash

SGLang

How to use tiny-random/glm-5.3-flash 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 "tiny-random/glm-5.3-flash" \
    --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": "tiny-random/glm-5.3-flash",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

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 "tiny-random/glm-5.3-flash" \
        --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": "tiny-random/glm-5.3-flash",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use tiny-random/glm-5.3-flash with Docker Model Runner:
```
docker model run hf.co/tiny-random/glm-5.3-flash
```

glm-5.3-flash / config.json

yujiepan

Upload folder using huggingface_hub

b3efe4a verified 3 days ago

Raw History Blame Contribute Delete

7.92 kB

{
  "architectures": [
    "Glm5NextForConditionalGeneration"
  ],
  "text_config": {
    "attention_bias": false,
    "attention_dropout": 0.0,
    "dtype": "bfloat16",
    "eos_token_id": [
      154820,
      154827,
      154829
    ],
    "first_k_dense_replace": 3,
    "hc_eps": 1e-06,
    "hc_mult": 4,
    "hc_sinkhorn_iters": 20,
    "head_dim": 0,
    "hidden_act": "silu",
    "hidden_size": 128,
    "index_head_dim": 128,
    "index_kpool": 4,
    "index_kpool_always_select_tail": true,
    "index_kpool_compress": true,
    "index_n_heads": 32,
    "index_topk": 128,
    "index_share_for_mtp_iteration": true,
    "indexer_rope_interleave": true,
    "indexer_types": [
      "full",
      "full",
      "full",
      "full",
      "full",
      "full",
      "full",
      "full"
    ],
    "initializer_range": 0.02,
    "intermediate_size": 128,
    "kv_lora_rank": 256,
    "layer_types": [
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "indexed_attention",
      "linear_attention",
      "linear_attention",
      "linear_attention",
      "indexed_attention"
    ],
    "linear_attn_config": {
      "num_heads": 64,
      "gate_lower_bound": -5.0,
      "head_dim": 128,
      "short_conv_kernel_size": 4,
      "kda_layers": [
        0,
        1,
        2,
        4,
        5,
        6
      ],
      "full_attn_layers": [
        3,
        7
      ]
    },
    "max_position_embeddings": 1048576,
    "mhc": true,
    "mla_use_nope": true,
    "mlp_layer_types": [
      "dense",
      "dense",
      "dense",
      "sparse",
      "sparse",
      "sparse",
      "sparse",
      "sparse"
    ],
    "model_type": "glm5_next_text",
    "moe_intermediate_size": 128,
    "moe_router_dtype": "float32",
    "n_group": 1,
    "n_routed_experts": 32,
    "n_shared_experts": 1,
    "norm_topk_prob": true,
    "num_attention_heads": 64,
    "num_experts_per_tok": 8,
    "num_hidden_layers": 8,
    "num_key_value_heads": 64,
    "num_nextn_predict_layers": 1,
    "output_router_logits": false,
    "pad_token_id": 154820,
    "q_lora_rank": 128,
    "qk_head_dim": 256,
    "qk_nope_head_dim": 32,
    "qk_rope_head_dim": 0,
    "rms_norm_eps": 1e-05,
    "routed_scaling_factor": 2.5,
    "router_aux_loss_coef": 0.001,
    "scoring_func": "sigmoid",
    "swiglu_limit": 10.0,
    "tie_word_embeddings": false,
    "topk_group": 1,
    "topk_method": "noaux_tc",
    "use_cache": true,
    "v_head_dim": 256,
    "vocab_size": 154880,
    "linear_head_dim": 128
  },
  "vision_config": {
    "attention_bias": true,
    "attention_dropout": 0.0,
    "depth": 2,
    "hidden_act": "silu",
    "hidden_size": 128,
    "image_size": 28,
    "in_channels": 3,
    "initializer_range": 0.02,
    "intermediate_size": 256,
    "model_type": "glm5_next_vision",
    "num_heads": 4,
    "out_hidden_size": 128,
    "patch_size": 14,
    "projection_intermediate_size": 64,
    "rms_norm_eps": 1e-05,
    "spatial_merge_size": 2,
    "swiglu_limit": 10.0,
    "temporal_patch_size": 2
  },
  "image_token_id": 154854,
  "video_token_id": 154855,
  "image_start_token_id": 154830,
  "image_end_token_id": 154831,
  "video_start_token_id": 154832,
  "video_end_token_id": 154833,
  "tie_word_embeddings": false,
  "model_type": "glm5_next",
  "transformers_version": "5.16.0",
  "quantization_config": {
    "activation_scheme": "dynamic",
    "fmt": "e4m3",
    "modules_to_not_convert": [
      "lm_head",
      "model.language_model.embed_tokens",
      "model.language_model.layers.0.self_attn.b_proj",
      "model.language_model.layers.0.self_attn.forget_gate.f_a_proj",
      "model.language_model.layers.0.self_attn.forget_gate.f_b_proj",
      "model.language_model.layers.0.self_attn.g_a_proj",
      "model.language_model.layers.0.self_attn.g_b_proj",
      "model.language_model.layers.0.self_attn.k_proj",
      "model.language_model.layers.0.self_attn.o_proj",
      "model.language_model.layers.0.self_attn.q_proj",
      "model.language_model.layers.0.self_attn.v_proj",
      "model.language_model.layers.1.self_attn.b_proj",
      "model.language_model.layers.1.self_attn.forget_gate.f_a_proj",
      "model.language_model.layers.1.self_attn.forget_gate.f_b_proj",
      "model.language_model.layers.1.self_attn.g_a_proj",
      "model.language_model.layers.1.self_attn.g_b_proj",
      "model.language_model.layers.1.self_attn.k_proj",
      "model.language_model.layers.1.self_attn.o_proj",
      "model.language_model.layers.1.self_attn.q_proj",
      "model.language_model.layers.1.self_attn.v_proj",
      "model.language_model.layers.2.self_attn.b_proj",
      "model.language_model.layers.2.self_attn.forget_gate.f_a_proj",
      "model.language_model.layers.2.self_attn.forget_gate.f_b_proj",
      "model.language_model.layers.2.self_attn.g_a_proj",
      "model.language_model.layers.2.self_attn.g_b_proj",
      "model.language_model.layers.2.self_attn.k_proj",
      "model.language_model.layers.2.self_attn.o_proj",
      "model.language_model.layers.2.self_attn.q_proj",
      "model.language_model.layers.2.self_attn.v_proj",
      "model.language_model.layers.3.self_attn.indexer.weights_proj",
      "model.language_model.layers.3.self_attn.indexer.wk",
      "model.language_model.layers.3.self_attn.indexer.wq_b",
      "model.language_model.layers.3.self_attn.kv_b_proj",
      "model.language_model.layers.4.self_attn.b_proj",
      "model.language_model.layers.4.self_attn.forget_gate.f_a_proj",
      "model.language_model.layers.4.self_attn.forget_gate.f_b_proj",
      "model.language_model.layers.4.self_attn.g_a_proj",
      "model.language_model.layers.4.self_attn.g_b_proj",
      "model.language_model.layers.4.self_attn.k_proj",
      "model.language_model.layers.4.self_attn.o_proj",
      "model.language_model.layers.4.self_attn.q_proj",
      "model.language_model.layers.4.self_attn.v_proj",
      "model.language_model.layers.5.self_attn.b_proj",
      "model.language_model.layers.5.self_attn.forget_gate.f_a_proj",
      "model.language_model.layers.5.self_attn.forget_gate.f_b_proj",
      "model.language_model.layers.5.self_attn.g_a_proj",
      "model.language_model.layers.5.self_attn.g_b_proj",
      "model.language_model.layers.5.self_attn.k_proj",
      "model.language_model.layers.5.self_attn.o_proj",
      "model.language_model.layers.5.self_attn.q_proj",
      "model.language_model.layers.5.self_attn.v_proj",
      "model.language_model.layers.6.self_attn.b_proj",
      "model.language_model.layers.6.self_attn.forget_gate.f_a_proj",
      "model.language_model.layers.6.self_attn.forget_gate.f_b_proj",
      "model.language_model.layers.6.self_attn.g_a_proj",
      "model.language_model.layers.6.self_attn.g_b_proj",
      "model.language_model.layers.6.self_attn.k_proj",
      "model.language_model.layers.6.self_attn.o_proj",
      "model.language_model.layers.6.self_attn.q_proj",
      "model.language_model.layers.6.self_attn.v_proj",
      "model.language_model.layers.7.self_attn.indexer.weights_proj",
      "model.language_model.layers.7.self_attn.indexer.wk",
      "model.language_model.layers.7.self_attn.indexer.wq_b",
      "model.language_model.layers.7.self_attn.kv_b_proj",
      "model.visual",
      "model.visual.blocks.0.attn.proj",
      "model.visual.blocks.0.attn.qkv",
      "model.visual.blocks.0.mlp.down_proj",
      "model.visual.blocks.0.mlp.gate_proj",
      "model.visual.blocks.0.mlp.up_proj",
      "model.visual.blocks.1.attn.proj",
      "model.visual.blocks.1.attn.qkv",
      "model.visual.blocks.1.mlp.down_proj",
      "model.visual.blocks.1.mlp.gate_proj",
      "model.visual.blocks.1.mlp.up_proj",
      "model.visual.merger.down_proj",
      "model.visual.merger.gate_proj",
      "model.visual.merger.proj",
      "model.visual.merger.up_proj"
    ],
    "quant_method": "fp8",
    "weight_block_size": [
      32,
      32
    ]
  }
}

	{
	"architectures": [
	"Glm5NextForConditionalGeneration"
	],
	"text_config": {
	"attention_bias": false,
	"attention_dropout": 0.0,
	"dtype": "bfloat16",
	"eos_token_id": [
	154820,
	154827,
	154829
	],
	"first_k_dense_replace": 3,
	"hc_eps": 1e-06,
	"hc_mult": 4,
	"hc_sinkhorn_iters": 20,
	"head_dim": 0,
	"hidden_act": "silu",
	"hidden_size": 128,
	"index_head_dim": 128,
	"index_kpool": 4,
	"index_kpool_always_select_tail": true,
	"index_kpool_compress": true,
	"index_n_heads": 32,
	"index_topk": 128,
	"index_share_for_mtp_iteration": true,
	"indexer_rope_interleave": true,
	"indexer_types": [
	"full",
	"full",
	"full",
	"full",
	"full",
	"full",
	"full",
	"full"
	],
	"initializer_range": 0.02,
	"intermediate_size": 128,
	"kv_lora_rank": 256,
	"layer_types": [
	"linear_attention",
	"linear_attention",
	"linear_attention",
	"indexed_attention",
	"linear_attention",
	"linear_attention",
	"linear_attention",
	"indexed_attention"
	],
	"linear_attn_config": {
	"num_heads": 64,
	"gate_lower_bound": -5.0,
	"head_dim": 128,
	"short_conv_kernel_size": 4,
	"kda_layers": [
	0,
	1,
	2,
	4,
	5,
	6
	],
	"full_attn_layers": [
	3,
	7
	]
	},
	"max_position_embeddings": 1048576,
	"mhc": true,
	"mla_use_nope": true,
	"mlp_layer_types": [
	"dense",
	"dense",
	"dense",
	"sparse",
	"sparse",
	"sparse",
	"sparse",
	"sparse"
	],
	"model_type": "glm5_next_text",
	"moe_intermediate_size": 128,
	"moe_router_dtype": "float32",
	"n_group": 1,
	"n_routed_experts": 32,
	"n_shared_experts": 1,
	"norm_topk_prob": true,
	"num_attention_heads": 64,
	"num_experts_per_tok": 8,
	"num_hidden_layers": 8,
	"num_key_value_heads": 64,
	"num_nextn_predict_layers": 1,
	"output_router_logits": false,
	"pad_token_id": 154820,
	"q_lora_rank": 128,
	"qk_head_dim": 256,
	"qk_nope_head_dim": 32,
	"qk_rope_head_dim": 0,
	"rms_norm_eps": 1e-05,
	"routed_scaling_factor": 2.5,
	"router_aux_loss_coef": 0.001,
	"scoring_func": "sigmoid",
	"swiglu_limit": 10.0,
	"tie_word_embeddings": false,
	"topk_group": 1,
	"topk_method": "noaux_tc",
	"use_cache": true,
	"v_head_dim": 256,
	"vocab_size": 154880,
	"linear_head_dim": 128
	},
	"vision_config": {
	"attention_bias": true,
	"attention_dropout": 0.0,
	"depth": 2,
	"hidden_act": "silu",
	"hidden_size": 128,
	"image_size": 28,
	"in_channels": 3,
	"initializer_range": 0.02,
	"intermediate_size": 256,
	"model_type": "glm5_next_vision",
	"num_heads": 4,
	"out_hidden_size": 128,
	"patch_size": 14,
	"projection_intermediate_size": 64,
	"rms_norm_eps": 1e-05,
	"spatial_merge_size": 2,
	"swiglu_limit": 10.0,
	"temporal_patch_size": 2
	},
	"image_token_id": 154854,
	"video_token_id": 154855,
	"image_start_token_id": 154830,
	"image_end_token_id": 154831,
	"video_start_token_id": 154832,
	"video_end_token_id": 154833,
	"tie_word_embeddings": false,
	"model_type": "glm5_next",
	"transformers_version": "5.16.0",
	"quantization_config": {
	"activation_scheme": "dynamic",
	"fmt": "e4m3",
	"modules_to_not_convert": [
	"lm_head",
	"model.language_model.embed_tokens",
	"model.language_model.layers.0.self_attn.b_proj",
	"model.language_model.layers.0.self_attn.forget_gate.f_a_proj",
	"model.language_model.layers.0.self_attn.forget_gate.f_b_proj",
	"model.language_model.layers.0.self_attn.g_a_proj",
	"model.language_model.layers.0.self_attn.g_b_proj",
	"model.language_model.layers.0.self_attn.k_proj",
	"model.language_model.layers.0.self_attn.o_proj",
	"model.language_model.layers.0.self_attn.q_proj",
	"model.language_model.layers.0.self_attn.v_proj",
	"model.language_model.layers.1.self_attn.b_proj",
	"model.language_model.layers.1.self_attn.forget_gate.f_a_proj",
	"model.language_model.layers.1.self_attn.forget_gate.f_b_proj",
	"model.language_model.layers.1.self_attn.g_a_proj",
	"model.language_model.layers.1.self_attn.g_b_proj",
	"model.language_model.layers.1.self_attn.k_proj",
	"model.language_model.layers.1.self_attn.o_proj",
	"model.language_model.layers.1.self_attn.q_proj",
	"model.language_model.layers.1.self_attn.v_proj",
	"model.language_model.layers.2.self_attn.b_proj",
	"model.language_model.layers.2.self_attn.forget_gate.f_a_proj",
	"model.language_model.layers.2.self_attn.forget_gate.f_b_proj",
	"model.language_model.layers.2.self_attn.g_a_proj",
	"model.language_model.layers.2.self_attn.g_b_proj",
	"model.language_model.layers.2.self_attn.k_proj",
	"model.language_model.layers.2.self_attn.o_proj",
	"model.language_model.layers.2.self_attn.q_proj",
	"model.language_model.layers.2.self_attn.v_proj",
	"model.language_model.layers.3.self_attn.indexer.weights_proj",
	"model.language_model.layers.3.self_attn.indexer.wk",
	"model.language_model.layers.3.self_attn.indexer.wq_b",
	"model.language_model.layers.3.self_attn.kv_b_proj",
	"model.language_model.layers.4.self_attn.b_proj",
	"model.language_model.layers.4.self_attn.forget_gate.f_a_proj",
	"model.language_model.layers.4.self_attn.forget_gate.f_b_proj",
	"model.language_model.layers.4.self_attn.g_a_proj",
	"model.language_model.layers.4.self_attn.g_b_proj",
	"model.language_model.layers.4.self_attn.k_proj",
	"model.language_model.layers.4.self_attn.o_proj",
	"model.language_model.layers.4.self_attn.q_proj",
	"model.language_model.layers.4.self_attn.v_proj",
	"model.language_model.layers.5.self_attn.b_proj",
	"model.language_model.layers.5.self_attn.forget_gate.f_a_proj",
	"model.language_model.layers.5.self_attn.forget_gate.f_b_proj",
	"model.language_model.layers.5.self_attn.g_a_proj",
	"model.language_model.layers.5.self_attn.g_b_proj",
	"model.language_model.layers.5.self_attn.k_proj",
	"model.language_model.layers.5.self_attn.o_proj",
	"model.language_model.layers.5.self_attn.q_proj",
	"model.language_model.layers.5.self_attn.v_proj",
	"model.language_model.layers.6.self_attn.b_proj",
	"model.language_model.layers.6.self_attn.forget_gate.f_a_proj",
	"model.language_model.layers.6.self_attn.forget_gate.f_b_proj",
	"model.language_model.layers.6.self_attn.g_a_proj",
	"model.language_model.layers.6.self_attn.g_b_proj",
	"model.language_model.layers.6.self_attn.k_proj",
	"model.language_model.layers.6.self_attn.o_proj",
	"model.language_model.layers.6.self_attn.q_proj",
	"model.language_model.layers.6.self_attn.v_proj",
	"model.language_model.layers.7.self_attn.indexer.weights_proj",
	"model.language_model.layers.7.self_attn.indexer.wk",
	"model.language_model.layers.7.self_attn.indexer.wq_b",
	"model.language_model.layers.7.self_attn.kv_b_proj",
	"model.visual",
	"model.visual.blocks.0.attn.proj",
	"model.visual.blocks.0.attn.qkv",
	"model.visual.blocks.0.mlp.down_proj",
	"model.visual.blocks.0.mlp.gate_proj",
	"model.visual.blocks.0.mlp.up_proj",
	"model.visual.blocks.1.attn.proj",
	"model.visual.blocks.1.attn.qkv",
	"model.visual.blocks.1.mlp.down_proj",
	"model.visual.blocks.1.mlp.gate_proj",
	"model.visual.blocks.1.mlp.up_proj",
	"model.visual.merger.down_proj",
	"model.visual.merger.gate_proj",
	"model.visual.merger.proj",
	"model.visual.merger.up_proj"
	],
	"quant_method": "fp8",
	"weight_block_size": [
	32,
	32
	]
	}
	}