Instructions to use jdopensource/JoyAI-LLM-Flash-INT8 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use jdopensource/JoyAI-LLM-Flash-INT8 with Transformers:

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

pipe = pipeline("text-generation", model="jdopensource/JoyAI-LLM-Flash-INT8", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("jdopensource/JoyAI-LLM-Flash-INT8", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use jdopensource/JoyAI-LLM-Flash-INT8 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "jdopensource/JoyAI-LLM-Flash-INT8"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "jdopensource/JoyAI-LLM-Flash-INT8",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/jdopensource/JoyAI-LLM-Flash-INT8

SGLang

How to use jdopensource/JoyAI-LLM-Flash-INT8 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 "jdopensource/JoyAI-LLM-Flash-INT8" \
    --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": "jdopensource/JoyAI-LLM-Flash-INT8",
		"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 "jdopensource/JoyAI-LLM-Flash-INT8" \
        --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": "jdopensource/JoyAI-LLM-Flash-INT8",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use jdopensource/JoyAI-LLM-Flash-INT8 with Docker Model Runner:
```
docker model run hf.co/jdopensource/JoyAI-LLM-Flash-INT8
```

JoyAI-LLM-Flash-INT8 / config.json

Mingke977

Add files using upload-large-folder tool

facd22e verified about 1 month ago

raw

history blame contribute delete

1.71 kB

	{
	"architectures": [
	"DeepseekV3ForCausalLM"
	],
	"attention_bias": false,
	"attention_dropout": 0.0,
	"auto_map": {
	"AutoConfig": "configuration_deepseek.DeepseekV3Config",
	"AutoModel": "modeling_deepseek.DeepseekV3Model",
	"AutoModelForCausalLM": "modeling_deepseek.DeepseekV3ForCausalLM"
	},
	"bos_token_id": 0,
	"dtype": "bfloat16",
	"eos_token_id": 1,
	"ep_size": 1,
	"first_k_dense_replace": 1,
	"head_dim": 64,
	"hidden_act": "silu",
	"hidden_size": 2048,
	"initializer_range": 0.02,
	"intermediate_size": 7168,
	"kv_lora_rank": 512,
	"max_position_embeddings": 131072,
	"model_type":"joyai_llm_flash",
	"moe_intermediate_size": 768,
	"moe_layer_freq": 1,
	"n_group": 1,
	"n_routed_experts": 256,
	"n_shared_experts": 1,
	"norm_topk_prob": true,
	"num_attention_heads": 32,
	"num_experts_per_tok": 8,
	"num_hidden_layers": 40,
	"num_key_value_heads": 32,
	"num_nextn_predict_layers": 1,
	"pretraining_tp": 1,
	"q_lora_rank": 1536,
	"qk_head_dim": 192,
	"qk_nope_head_dim": 128,
	"qk_rope_head_dim": 64,
	"quantization_config": {
	"activation_scheme": "dynamic",
	"quant_method": "blockwise_int8",
	"weight_block_size": [
	128,
	128
	]
	},
	"rms_norm_eps": 1e-06,
	"rope_interleave": true,
	"rope_scaling": null,
	"rope_theta": 32000000,
	"routed_scaling_factor": 2.5,
	"scoring_func": "sigmoid",
	"tie_word_embeddings": false,
	"topk_group": 1,
	"topk_method": "noaux_tc",
	"transformers_version": "4.57.3",
	"use_cache": true,
	"v_head_dim": 128,
	"vocab_size": 129280
	}