Image-Text-to-Text
Transformers
Safetensors
inkling_mm_model
conversational
audio-text-to-text
Mixture of Experts
8-bit precision
Instructions to use thinkingmachines/Inkling-NVFP4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use thinkingmachines/Inkling-NVFP4 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("image-text-to-text", model="thinkingmachines/Inkling-NVFP4") 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)# Load model directly from transformers import AutoProcessor, AutoModelForMultimodalLM processor = AutoProcessor.from_pretrained("thinkingmachines/Inkling-NVFP4") model = AutoModelForMultimodalLM.from_pretrained("thinkingmachines/Inkling-NVFP4") 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=40) print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:])) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use thinkingmachines/Inkling-NVFP4 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "thinkingmachines/Inkling-NVFP4" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "thinkingmachines/Inkling-NVFP4", "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/thinkingmachines/Inkling-NVFP4
- SGLang
How to use thinkingmachines/Inkling-NVFP4 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 "thinkingmachines/Inkling-NVFP4" \ --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": "thinkingmachines/Inkling-NVFP4", "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 "thinkingmachines/Inkling-NVFP4" \ --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": "thinkingmachines/Inkling-NVFP4", "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 thinkingmachines/Inkling-NVFP4 with Docker Model Runner:
docker model run hf.co/thinkingmachines/Inkling-NVFP4
Transformers processor and chat template support
Browse files- chat_template.jinja +129 -0
- processor_config.json +46 -0
chat_template.jinja
ADDED
|
@@ -0,0 +1,129 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- set effort_map = {"none": 0.0, "minimal": 0.1, "low": 0.2, "medium": 0.7, "high": 0.9, "max": 0.99} -%}
|
| 2 |
+
{%- set role_token = {"user": "<|message_user|>", "assistant": "<|message_model|>", "system": "<|message_system|>", "tool": "<|message_tool|>"} -%}
|
| 3 |
+
|
| 4 |
+
{%- macro emit_thinking_effort() -%}
|
| 5 |
+
{%- set eff = reasoning_effort if reasoning_effort is defined and reasoning_effort is not none else 0.9 -%}
|
| 6 |
+
{%- if eff is string -%}
|
| 7 |
+
{%- set key = eff | trim -%}
|
| 8 |
+
{%- if key not in effort_map -%}
|
| 9 |
+
{{- raise_exception("Unknown reasoning_effort: " ~ eff) -}}
|
| 10 |
+
{%- endif -%}
|
| 11 |
+
{%- set num = effort_map[key] -%}
|
| 12 |
+
{%- else -%}
|
| 13 |
+
{%- set num = eff | float -%}
|
| 14 |
+
{%- endif -%}
|
| 15 |
+
{%- if num < 0.0 or num > 0.99 -%}
|
| 16 |
+
{{- raise_exception("reasoning_effort must be in [0.0, 0.99]") -}}
|
| 17 |
+
{%- endif -%}
|
| 18 |
+
{{- "<|message_system|><|content_text|>Thinking effort level: " -}}
|
| 19 |
+
{%- if num == 0.0 -%}0{%- else -%}{{ num }}{%- endif -%}
|
| 20 |
+
{{- "<|end_message|>" -}}
|
| 21 |
+
{%- endmacro -%}
|
| 22 |
+
|
| 23 |
+
{%- if tools -%}
|
| 24 |
+
{%- set tool_state = namespace(specs=[]) -%}
|
| 25 |
+
{%- for tool in tools -%}
|
| 26 |
+
{%- set fn = tool.function if tool.function is defined else tool -%}
|
| 27 |
+
{%- set spec = {
|
| 28 |
+
"description": (fn.description if fn.description is defined and fn.description else ""),
|
| 29 |
+
"name": fn.name,
|
| 30 |
+
"parameters": (fn.parameters if fn.parameters is defined and fn.parameters else {}),
|
| 31 |
+
"type": (tool.type if tool.type is defined and tool.type else "function"),
|
| 32 |
+
} -%}
|
| 33 |
+
{%- set tool_state.specs = tool_state.specs + [spec] -%}
|
| 34 |
+
{%- endfor -%}
|
| 35 |
+
{{- "<|message_system|>tool_declare<|content_xml|>" -}}
|
| 36 |
+
{{- tool_state.specs | tojson(sort_keys=true, separators=(",", ":")) -}}
|
| 37 |
+
{{- "<|end_message|>" -}}
|
| 38 |
+
{%- endif -%}
|
| 39 |
+
|
| 40 |
+
{%- set state = namespace(effort_emitted=false) -%}
|
| 41 |
+
{%- for message in messages -%}
|
| 42 |
+
{%- if message.role not in role_token -%}
|
| 43 |
+
{{- raise_exception("Unknown message role: " ~ message.role) -}}
|
| 44 |
+
{%- endif -%}
|
| 45 |
+
{%- if not state.effort_emitted and message.role != "system" -%}
|
| 46 |
+
{{- emit_thinking_effort() -}}
|
| 47 |
+
{%- set state.effort_emitted = true -%}
|
| 48 |
+
{%- endif -%}
|
| 49 |
+
|
| 50 |
+
{%- set rtok = role_token[message.role] -%}
|
| 51 |
+
|
| 52 |
+
{%- if message.role == "tool" -%}
|
| 53 |
+
{%- set tool_name_state = namespace(name="") -%}
|
| 54 |
+
{%- if message.name is defined and message.name -%}
|
| 55 |
+
{%- set tool_name_state.name = message.name -%}
|
| 56 |
+
{%- elif message.tool_call_id is defined and message.tool_call_id -%}
|
| 57 |
+
{%- for prev in messages -%}
|
| 58 |
+
{%- if prev.role == "assistant" and prev.tool_calls -%}
|
| 59 |
+
{%- for tc in prev.tool_calls -%}
|
| 60 |
+
{%- if tc.id is defined and tc.id == message.tool_call_id and tc.function.name is defined -%}
|
| 61 |
+
{%- set tool_name_state.name = tc.function.name -%}
|
| 62 |
+
{%- endif -%}
|
| 63 |
+
{%- endfor -%}
|
| 64 |
+
{%- endif -%}
|
| 65 |
+
{%- endfor -%}
|
| 66 |
+
{%- endif -%}
|
| 67 |
+
{{- rtok -}}
|
| 68 |
+
{%- if tool_name_state.name -%}{{- tool_name_state.name -}}{%- endif -%}
|
| 69 |
+
{{- "<|content_text|>" -}}
|
| 70 |
+
{%- if message.content is string -%}{{- message.content -}}{%- endif -%}
|
| 71 |
+
{{- "<|end_message|>" -}}
|
| 72 |
+
|
| 73 |
+
{%- else -%}
|
| 74 |
+
{%- if message.role == "assistant" and message.reasoning_content is defined and message.reasoning_content -%}
|
| 75 |
+
{{- "<|message_model|><|content_thinking|>" ~ message.reasoning_content ~ "<|end_message|>" -}}
|
| 76 |
+
{%- endif -%}
|
| 77 |
+
|
| 78 |
+
{%- if message.content is string -%}
|
| 79 |
+
{{- rtok ~ "<|content_text|>" ~ message.content ~ "<|end_message|>" -}}
|
| 80 |
+
{%- elif message.content -%}
|
| 81 |
+
{%- for part in message.content -%}
|
| 82 |
+
{%- if part is string -%}
|
| 83 |
+
{{- rtok ~ "<|content_text|>" ~ part ~ "<|end_message|>" -}}
|
| 84 |
+
{%- elif part.type is not defined or part.type in ("text", "input_text") -%}
|
| 85 |
+
{%- set text_part = (part.text if part.text is defined and part.text is string else "") -%}
|
| 86 |
+
{{- rtok ~ "<|content_text|>" ~ text_part ~ "<|end_message|>" -}}
|
| 87 |
+
{%- elif part.type in ("image", "input_image", "image_url") -%}
|
| 88 |
+
{{- rtok ~ "<|content_image|><|unused_200054|><|end_message|>" -}}
|
| 89 |
+
{%- elif part.type in ("audio", "input_audio", "audio_url") -%}
|
| 90 |
+
{{- rtok ~ "<|content_audio_input|><|unused_200053|><|audio_end|><|end_message|>" -}}
|
| 91 |
+
{%- else -%}
|
| 92 |
+
{{- raise_exception("Unsupported content part type: " ~ part.type) -}}
|
| 93 |
+
{%- endif -%}
|
| 94 |
+
{%- endfor -%}
|
| 95 |
+
{%- endif -%}
|
| 96 |
+
|
| 97 |
+
{%- if message.role == "assistant" and message.tool_calls -%}
|
| 98 |
+
{%- for tc in message.tool_calls -%}
|
| 99 |
+
{%- set fn = tc.function -%}
|
| 100 |
+
{%- if fn.name is not defined or fn.name is not string -%}
|
| 101 |
+
{{- raise_exception("tool call function name must be a string") -}}
|
| 102 |
+
{%- endif -%}
|
| 103 |
+
{%- set args = fn.arguments if fn.arguments is defined and fn.arguments else {} -%}
|
| 104 |
+
{%- if args is string -%}
|
| 105 |
+
{{- raise_exception("tool call arguments must be a parsed object, not a JSON string; canonicalize upstream") -}}
|
| 106 |
+
{%- endif -%}
|
| 107 |
+
{%- if args is not mapping -%}
|
| 108 |
+
{{- raise_exception("tool call arguments must be an object") -}}
|
| 109 |
+
{%- endif -%}
|
| 110 |
+
{{- "<|message_model|>" ~ fn.name ~ "<|content_invoke_tool_json|>" -}}
|
| 111 |
+
{{- '{"name":' ~ (fn.name | tojson(sort_keys=true, separators=(",", ":"))) ~ ',"args":' -}}
|
| 112 |
+
{{- (args | tojson(sort_keys=true, separators=(",", ":"))) -}}
|
| 113 |
+
{{- "}<|end_message|>" -}}
|
| 114 |
+
{%- endfor -%}
|
| 115 |
+
{%- endif -%}
|
| 116 |
+
|
| 117 |
+
{%- if message.role == "assistant" -%}
|
| 118 |
+
{{- "<|content_model_end_sampling|>" -}}
|
| 119 |
+
{%- endif -%}
|
| 120 |
+
{%- endif -%}
|
| 121 |
+
{%- endfor -%}
|
| 122 |
+
|
| 123 |
+
{%- if not state.effort_emitted -%}
|
| 124 |
+
{{- emit_thinking_effort() -}}
|
| 125 |
+
{%- endif -%}
|
| 126 |
+
|
| 127 |
+
{%- if add_generation_prompt -%}
|
| 128 |
+
{{- "<|message_model|>" -}}
|
| 129 |
+
{%- endif -%}
|
processor_config.json
ADDED
|
@@ -0,0 +1,46 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"audio_token": "<|unused_200053|>",
|
| 3 |
+
"audio_bos_token": "<|content_audio_input|>",
|
| 4 |
+
"dmel_max_value": 2.0,
|
| 5 |
+
"dmel_min_value": -7.0,
|
| 6 |
+
"feature_extractor": {
|
| 7 |
+
"audio_token_duration_s": 0.05,
|
| 8 |
+
"feature_extractor_type": "InklingFeatureExtractor",
|
| 9 |
+
"feature_size": 80,
|
| 10 |
+
"hop_length": 800,
|
| 11 |
+
"n_fft": 1600,
|
| 12 |
+
"padding_side": "right",
|
| 13 |
+
"padding_value": 0.0,
|
| 14 |
+
"return_attention_mask": true,
|
| 15 |
+
"sampling_rate": 16000,
|
| 16 |
+
"window_size": 1600,
|
| 17 |
+
"window_size_multiplier": 2.0
|
| 18 |
+
},
|
| 19 |
+
"image_processor": {
|
| 20 |
+
"do_convert_rgb": true,
|
| 21 |
+
"do_normalize": true,
|
| 22 |
+
"do_rescale": true,
|
| 23 |
+
"do_resize": true,
|
| 24 |
+
"image_mean": [
|
| 25 |
+
0.48145466,
|
| 26 |
+
0.4578275,
|
| 27 |
+
0.40821073
|
| 28 |
+
],
|
| 29 |
+
"image_processor_type": "InklingImageProcessor",
|
| 30 |
+
"image_std": [
|
| 31 |
+
0.26862954,
|
| 32 |
+
0.26130258,
|
| 33 |
+
0.27577711
|
| 34 |
+
],
|
| 35 |
+
"resample": 3,
|
| 36 |
+
"rescale_factor": 0.00392156862745098,
|
| 37 |
+
"size": {
|
| 38 |
+
"height": 40,
|
| 39 |
+
"width": 40
|
| 40 |
+
}
|
| 41 |
+
},
|
| 42 |
+
"image_token": "<|unused_200054|>",
|
| 43 |
+
"image_bos_token": "<|content_image|>",
|
| 44 |
+
"num_dmel_bins": 16,
|
| 45 |
+
"processor_class": "InklingProcessor"
|
| 46 |
+
}
|