Text Generation
Transformers
Safetensors
English
gemma4_unified
image-text-to-text
gemma4
lora
character
roleplay
int8
w8a8
compressed-tensors
vllm
tpu
MDRG
Factorial-Omega
conversational
8-bit precision
Instructions to use efficiencyx/Jun-LoRA-12B-Safetensor-INT8 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use efficiencyx/Jun-LoRA-12B-Safetensor-INT8 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="efficiencyx/Jun-LoRA-12B-Safetensor-INT8") 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("efficiencyx/Jun-LoRA-12B-Safetensor-INT8") model = AutoModelForMultimodalLM.from_pretrained("efficiencyx/Jun-LoRA-12B-Safetensor-INT8", 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 efficiencyx/Jun-LoRA-12B-Safetensor-INT8 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "efficiencyx/Jun-LoRA-12B-Safetensor-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": "efficiencyx/Jun-LoRA-12B-Safetensor-INT8", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/efficiencyx/Jun-LoRA-12B-Safetensor-INT8
- SGLang
How to use efficiencyx/Jun-LoRA-12B-Safetensor-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 "efficiencyx/Jun-LoRA-12B-Safetensor-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": "efficiencyx/Jun-LoRA-12B-Safetensor-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 "efficiencyx/Jun-LoRA-12B-Safetensor-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": "efficiencyx/Jun-LoRA-12B-Safetensor-INT8", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use efficiencyx/Jun-LoRA-12B-Safetensor-INT8 with Docker Model Runner:
docker model run hf.co/efficiencyx/Jun-LoRA-12B-Safetensor-INT8
config.json in the v7 layout: transformers 5.17 wrote per_layer_config, which vLLM 0.30 cannot read
Browse files- config.json +59 -92
config.json
CHANGED
|
@@ -7,7 +7,7 @@
|
|
| 7 |
"architectures": null,
|
| 8 |
"audio_embed_dim": 640,
|
| 9 |
"chunk_size_feed_forward": 0,
|
| 10 |
-
"
|
| 11 |
"id2label": {
|
| 12 |
"0": "LABEL_0",
|
| 13 |
"1": "LABEL_1"
|
|
@@ -28,7 +28,7 @@
|
|
| 28 |
"audio_token_id": 258881,
|
| 29 |
"boa_token_id": 256000,
|
| 30 |
"boi_token_id": 255999,
|
| 31 |
-
"
|
| 32 |
"eoa_token_index": 258883,
|
| 33 |
"eoi_token_id": 258882,
|
| 34 |
"eos_token_id": [
|
|
@@ -38,68 +38,16 @@
|
|
| 38 |
"image_token_id": 258880,
|
| 39 |
"initializer_range": 0.02,
|
| 40 |
"model_type": "gemma4_unified",
|
| 41 |
-
"quantization_config": {
|
| 42 |
-
"config_groups": {
|
| 43 |
-
"group_0": {
|
| 44 |
-
"format": "int-quantized",
|
| 45 |
-
"input_activations": {
|
| 46 |
-
"actorder": null,
|
| 47 |
-
"block_structure": null,
|
| 48 |
-
"dynamic": true,
|
| 49 |
-
"group_size": null,
|
| 50 |
-
"num_bits": 8,
|
| 51 |
-
"observer": null,
|
| 52 |
-
"observer_kwargs": {},
|
| 53 |
-
"scale_dtype": null,
|
| 54 |
-
"strategy": "token",
|
| 55 |
-
"symmetric": true,
|
| 56 |
-
"type": "int",
|
| 57 |
-
"zp_dtype": null
|
| 58 |
-
},
|
| 59 |
-
"output_activations": null,
|
| 60 |
-
"targets": [
|
| 61 |
-
"Linear"
|
| 62 |
-
],
|
| 63 |
-
"weights": {
|
| 64 |
-
"actorder": null,
|
| 65 |
-
"block_structure": null,
|
| 66 |
-
"dynamic": false,
|
| 67 |
-
"group_size": null,
|
| 68 |
-
"num_bits": 8,
|
| 69 |
-
"observer": "memoryless_minmax",
|
| 70 |
-
"observer_kwargs": {},
|
| 71 |
-
"scale_dtype": null,
|
| 72 |
-
"strategy": "channel",
|
| 73 |
-
"symmetric": true,
|
| 74 |
-
"type": "int",
|
| 75 |
-
"zp_dtype": null
|
| 76 |
-
}
|
| 77 |
-
}
|
| 78 |
-
},
|
| 79 |
-
"format": "int-quantized",
|
| 80 |
-
"global_compression_ratio": null,
|
| 81 |
-
"ignore": [
|
| 82 |
-
"model.vision_embedder.patch_dense",
|
| 83 |
-
"model.embed_vision.embedding_projection",
|
| 84 |
-
"model.embed_audio.embedding_projection",
|
| 85 |
-
"lm_head"
|
| 86 |
-
],
|
| 87 |
-
"kv_cache_scheme": null,
|
| 88 |
-
"quant_method": "compressed-tensors",
|
| 89 |
-
"quantization_status": "compressed",
|
| 90 |
-
"sparsity_config": {},
|
| 91 |
-
"transform_config": {},
|
| 92 |
-
"version": "0.19.0"
|
| 93 |
-
},
|
| 94 |
"text_config": {
|
| 95 |
"attention_bias": false,
|
| 96 |
"attention_dropout": 0.0,
|
| 97 |
"attention_k_eq_v": true,
|
| 98 |
"bos_token_id": 2,
|
| 99 |
-
"
|
| 100 |
"enable_moe_block": false,
|
| 101 |
"eos_token_id": 1,
|
| 102 |
"final_logit_softcapping": 30.0,
|
|
|
|
| 103 |
"head_dim": 256,
|
| 104 |
"hidden_activation": "gelu_pytorch_tanh",
|
| 105 |
"hidden_size": 3840,
|
|
@@ -161,44 +109,11 @@
|
|
| 161 |
"moe_intermediate_size": null,
|
| 162 |
"num_attention_heads": 16,
|
| 163 |
"num_experts": null,
|
|
|
|
| 164 |
"num_hidden_layers": 48,
|
| 165 |
"num_key_value_heads": 8,
|
| 166 |
"num_kv_shared_layers": 0,
|
| 167 |
"pad_token_id": 0,
|
| 168 |
-
"per_layer_config": {
|
| 169 |
-
"05": {
|
| 170 |
-
"head_dim": 512,
|
| 171 |
-
"num_key_value_heads": 1
|
| 172 |
-
},
|
| 173 |
-
"11": {
|
| 174 |
-
"head_dim": 512,
|
| 175 |
-
"num_key_value_heads": 1
|
| 176 |
-
},
|
| 177 |
-
"17": {
|
| 178 |
-
"head_dim": 512,
|
| 179 |
-
"num_key_value_heads": 1
|
| 180 |
-
},
|
| 181 |
-
"23": {
|
| 182 |
-
"head_dim": 512,
|
| 183 |
-
"num_key_value_heads": 1
|
| 184 |
-
},
|
| 185 |
-
"29": {
|
| 186 |
-
"head_dim": 512,
|
| 187 |
-
"num_key_value_heads": 1
|
| 188 |
-
},
|
| 189 |
-
"35": {
|
| 190 |
-
"head_dim": 512,
|
| 191 |
-
"num_key_value_heads": 1
|
| 192 |
-
},
|
| 193 |
-
"41": {
|
| 194 |
-
"head_dim": 512,
|
| 195 |
-
"num_key_value_heads": 1
|
| 196 |
-
},
|
| 197 |
-
"47": {
|
| 198 |
-
"head_dim": 512,
|
| 199 |
-
"num_key_value_heads": 1
|
| 200 |
-
}
|
| 201 |
-
},
|
| 202 |
"rms_norm_eps": 1e-06,
|
| 203 |
"rope_parameters": {
|
| 204 |
"full_attention": {
|
|
@@ -221,13 +136,12 @@
|
|
| 221 |
"vocab_size_per_layer_input": 0
|
| 222 |
},
|
| 223 |
"tie_word_embeddings": true,
|
| 224 |
-
"transformers_version": "5.17.0",
|
| 225 |
"video_token_id": 258884,
|
| 226 |
"vision_config": {
|
| 227 |
"_name_or_path": "",
|
| 228 |
"architectures": null,
|
| 229 |
"chunk_size_feed_forward": 0,
|
| 230 |
-
"
|
| 231 |
"id2label": {
|
| 232 |
"0": "LABEL_0",
|
| 233 |
"1": "LABEL_1"
|
|
@@ -249,5 +163,58 @@
|
|
| 249 |
"problem_type": null,
|
| 250 |
"return_dict": true,
|
| 251 |
"rms_norm_eps": 1e-06
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 252 |
}
|
| 253 |
}
|
|
|
|
| 7 |
"architectures": null,
|
| 8 |
"audio_embed_dim": 640,
|
| 9 |
"chunk_size_feed_forward": 0,
|
| 10 |
+
"torch_dtype": "bfloat16",
|
| 11 |
"id2label": {
|
| 12 |
"0": "LABEL_0",
|
| 13 |
"1": "LABEL_1"
|
|
|
|
| 28 |
"audio_token_id": 258881,
|
| 29 |
"boa_token_id": 256000,
|
| 30 |
"boi_token_id": 255999,
|
| 31 |
+
"torch_dtype": "bfloat16",
|
| 32 |
"eoa_token_index": 258883,
|
| 33 |
"eoi_token_id": 258882,
|
| 34 |
"eos_token_id": [
|
|
|
|
| 38 |
"image_token_id": 258880,
|
| 39 |
"initializer_range": 0.02,
|
| 40 |
"model_type": "gemma4_unified",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 41 |
"text_config": {
|
| 42 |
"attention_bias": false,
|
| 43 |
"attention_dropout": 0.0,
|
| 44 |
"attention_k_eq_v": true,
|
| 45 |
"bos_token_id": 2,
|
| 46 |
+
"torch_dtype": "bfloat16",
|
| 47 |
"enable_moe_block": false,
|
| 48 |
"eos_token_id": 1,
|
| 49 |
"final_logit_softcapping": 30.0,
|
| 50 |
+
"global_head_dim": 512,
|
| 51 |
"head_dim": 256,
|
| 52 |
"hidden_activation": "gelu_pytorch_tanh",
|
| 53 |
"hidden_size": 3840,
|
|
|
|
| 109 |
"moe_intermediate_size": null,
|
| 110 |
"num_attention_heads": 16,
|
| 111 |
"num_experts": null,
|
| 112 |
+
"num_global_key_value_heads": 1,
|
| 113 |
"num_hidden_layers": 48,
|
| 114 |
"num_key_value_heads": 8,
|
| 115 |
"num_kv_shared_layers": 0,
|
| 116 |
"pad_token_id": 0,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 117 |
"rms_norm_eps": 1e-06,
|
| 118 |
"rope_parameters": {
|
| 119 |
"full_attention": {
|
|
|
|
| 136 |
"vocab_size_per_layer_input": 0
|
| 137 |
},
|
| 138 |
"tie_word_embeddings": true,
|
|
|
|
| 139 |
"video_token_id": 258884,
|
| 140 |
"vision_config": {
|
| 141 |
"_name_or_path": "",
|
| 142 |
"architectures": null,
|
| 143 |
"chunk_size_feed_forward": 0,
|
| 144 |
+
"torch_dtype": "bfloat16",
|
| 145 |
"id2label": {
|
| 146 |
"0": "LABEL_0",
|
| 147 |
"1": "LABEL_1"
|
|
|
|
| 163 |
"problem_type": null,
|
| 164 |
"return_dict": true,
|
| 165 |
"rms_norm_eps": 1e-06
|
| 166 |
+
},
|
| 167 |
+
"quantization_config": {
|
| 168 |
+
"config_groups": {
|
| 169 |
+
"group_0": {
|
| 170 |
+
"format": "int-quantized",
|
| 171 |
+
"input_activations": {
|
| 172 |
+
"actorder": null,
|
| 173 |
+
"block_structure": null,
|
| 174 |
+
"dynamic": true,
|
| 175 |
+
"group_size": null,
|
| 176 |
+
"num_bits": 8,
|
| 177 |
+
"observer": null,
|
| 178 |
+
"observer_kwargs": {},
|
| 179 |
+
"scale_dtype": null,
|
| 180 |
+
"strategy": "token",
|
| 181 |
+
"symmetric": true,
|
| 182 |
+
"type": "int",
|
| 183 |
+
"zp_dtype": null
|
| 184 |
+
},
|
| 185 |
+
"output_activations": null,
|
| 186 |
+
"targets": [
|
| 187 |
+
"Linear"
|
| 188 |
+
],
|
| 189 |
+
"weights": {
|
| 190 |
+
"actorder": null,
|
| 191 |
+
"block_structure": null,
|
| 192 |
+
"dynamic": false,
|
| 193 |
+
"group_size": null,
|
| 194 |
+
"num_bits": 8,
|
| 195 |
+
"observer": "memoryless_minmax",
|
| 196 |
+
"observer_kwargs": {},
|
| 197 |
+
"scale_dtype": null,
|
| 198 |
+
"strategy": "channel",
|
| 199 |
+
"symmetric": true,
|
| 200 |
+
"type": "int",
|
| 201 |
+
"zp_dtype": null
|
| 202 |
+
}
|
| 203 |
+
}
|
| 204 |
+
},
|
| 205 |
+
"format": "int-quantized",
|
| 206 |
+
"global_compression_ratio": null,
|
| 207 |
+
"ignore": [
|
| 208 |
+
"model.vision_embedder.patch_dense",
|
| 209 |
+
"model.embed_vision.embedding_projection",
|
| 210 |
+
"model.embed_audio.embedding_projection",
|
| 211 |
+
"lm_head"
|
| 212 |
+
],
|
| 213 |
+
"kv_cache_scheme": null,
|
| 214 |
+
"quant_method": "compressed-tensors",
|
| 215 |
+
"quantization_status": "compressed",
|
| 216 |
+
"sparsity_config": {},
|
| 217 |
+
"transform_config": {},
|
| 218 |
+
"version": "0.19.0"
|
| 219 |
}
|
| 220 |
}
|