Upload NuExtract3-mlx-5bits

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,28 @@
+---
+library_name: mlx
+base_model: numind/NuExtract3
+base_model_relation: quantized
+tags:
+- mlx
+- mlx-vlm
+- quantized
+---
+
+# numind/NuExtract3-mlx-5bits
+
+MLX quantized version of `numind/NuExtract3`.
+
+## Quantization
+
+- mode: `affine`
+- group_size: `64`
+- bits: `5`
+- dtype: `source/default`
+- quant_predicate: `none`
+
+## Usage
+
+```bash
+pip install -U mlx-vlm
+mlx_vlm.generate --model numind/NuExtract3-mlx-5bits --prompt "your prompt"
+```

chat_template.jinja

@@ -0,0 +1,139 @@
+{%- if not messages %}
+    {{- raise_exception('No messages provided.') }}
+{%- endif %}
+{%- set image_count = namespace(value=0) %}
+{%- set image_placeholder = '<|vision_start|><|image_pad|><|vision_end|>' -%}
+{%- set mode = mode | default('content') -%}
+{%- if template -%}{%- set mode = 'structured' -%}{%- endif -%}
+{%- if not template and mode == 'structured' %}
+    {{- raise_exception('`structured` mode specified but no `template` provided.') }}
+{%- endif %}
+{%- if mode not in ['structured', 'content', 'template-generation', 'document-detection', 'markdown'] -%}{%- set mode = 'content' -%}{%- endif -%}
+{%- if mode == 'markdown' %}{%- set mode = 'content' -%}{%- endif %}
+{%- set enable_thinking = enable_thinking | default(False) -%}
+{%- if mode not in ['structured', 'content'] and enable_thinking %}
+    {{- raise_exception('`enable_thinking` can only be `True` for `structured` and `content` modes.') }}
+{%- endif %}
+{%- set has_examples = namespace(flag=false) -%}
+{%- if mode != 'structured' -%}{%- set has_examples = false -%}{%- endif -%}
+{# MACRO TO RENDER MESSAGE CONTENT #}
+{%- macro render_content(content, do_vision_count, is_system_content=false) %}
+    {%- if content is string %}
+        {{- content }}
+    {%- elif content is iterable and content is not mapping %}
+        {%- for item in content %}
+            {%- if 'image' in item or 'image_url' in item or item.type == 'image' %}
+                {%- if is_system_content %}
+                    {{- raise_exception('System message cannot contain images.') }}
+                {%- endif %}
+                {%- if do_vision_count %}
+                    {%- set image_count.value = image_count.value + 1 %}
+                {%- endif %}
+                {%- if add_vision_id %}
+                    {{- 'Picture ' ~ image_count.value ~ ': ' }}
+                {%- endif %}
+                {{- '<|vision_start|><|image_pad|><|vision_end|>\n' }}
+            {%- elif 'text' in item %}
+                {{- item.text + '\n' }}
+            {%- else %}
+                {{- raise_exception('Unexpected item type in content.') }}
+            {%- endif %}
+        {%- endfor %}
+    {%- elif content is none or content is undefined %}
+        {{- '' }}
+    {%- else %}
+        {{- raise_exception('Unexpected content type.') }}
+    {%- endif %}
+{%- endmacro %}
+{# SYSTEM MESSAGE #}
+{%- if messages[0].role == 'system' %}
+    {%- set content = render_content(messages[0].content, false, true)|trim %}
+    {{- '<|im_start|>system\n' + content + '<|im_end|>\n' }}
+{%- endif %}
+{# USER MESSAGE #}
+{{- '<|im_start|>user\n' -}}
+{{- '【task】' + mode|replace("-", " ") + '\n' -}}
+{# Template Section (for structured task): specifies template, instructions, examples, previous_output #}
+{%- if mode == 'structured' -%}
+    {{- '【template_start】' + template + '【template_end】\n' -}}
+    {# Instructions Section #}
+    {%- if instructions -%}
+        {{- '【instructions_start】' + instructions + '【instructions_end】\n'-}}
+    {%- endif -%}
+    {# Examples Section (only for extraction tasks) #}
+    {%- for message in messages -%}
+        {%- if message.role == 'developer' and 'content' in message -%}
+            {# Validate that there is at least one input and one output contents #}
+            {%- set example_inputs = message.content[:-1] -%}
+            {%- set example_output_part = message.content[-1] -%}
+            {%- if example_inputs|length > 0 -%}
+                {%- if not has_examples.flag -%}
+                    {{- '【examples_start】\n' -}}
+                    {%- set has_examples.flag = true -%}
+                {%- endif -%}
+                {{- '【example_input_start】' + render_content(example_inputs, true)|trim + '【example_input_end】\n' -}}
+                {# Example output: only keep the text of the first output content #}
+                {%- set output_text = '' -%}
+                {%- if example_output_part is string -%}
+                    {%- set output_text = example_output_part -%}
+                {%- elif example_output_part.text is defined -%}
+                    {%- set output_text = example_output_part.text -%}
+                {%- endif -%}
+                {{- '【example_output_start】' + output_text|trim + '【example_output_end】\n' -}}
+                {%- if loop.last and has_examples.flag -%}
+                    {{- '【examples_end】\n' -}}
+                {%- endif -%}
+            {%- endif -%}
+        {%- endif -%}
+    {%- endfor -%}
+    {# Previous Output Section #}
+    {%- if previous_output -%}
+        {{- '【previous_output_start】' + previous_output + '【previous_output_end】\n' -}}
+    {%- endif -%}
+{%- endif -%}
+{{- '【document_start】\n' -}}
+{# PROCESS PROVIDED USER MESSAGES (RENDERED INTO A SINGLE ONE) #}
+{%- for message in messages -%}
+    {%- if message.role == "system" %}
+        {%- if not loop.first %}
+            {{- raise_exception('System message must be at the beginning.') }}
+        {%- endif %}
+    {%- elif message.role == 'user' and message.name != "example" -%}
+        {%- set content = render_content(message.content, true)|trim %}
+        {{- content + '\n' -}}
+    {# {%- elif message.role == 'assistant' and not loop.last %}
+        llama.cpp renders a synthetic init example with an assistant turn in
+        the middle; ignore it so valid NuExtract prompts render unchanged.
+        {{- raise_exception('Assistant message must be at the end.') }} #}
+    {%- endif %}
+{%- endfor -%}
+{{- '【document_end】<|im_end|>\n' -}}
+{# ASSISTANT MESSAGE #}
+{%- if messages[-1].role == 'assistant' %}
+    {%- if add_generation_prompt -%}
+        {{- raise_exception('`add_generation_prompt` can only be `True` when no assistant message is provided.') }}
+    {%- endif %}
+    {%- set content = render_content(messages[-1].content, true)|trim %}
+    {%- set reasoning_content = '' %}
+    {%- if messages[-1].reasoning_content is string %}
+        {%- set reasoning_content = messages[-1].reasoning_content %}
+    {%- else %}
+        {%- if '</think>' in content %}
+            {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+            {%- set content = content.split('</think>')[-1].lstrip('\n') %}
+        {%- endif %}
+    {%- endif %}
+    {%- set reasoning_content = reasoning_content|trim %}
+    {% generation %}
+    {{- '<|im_start|>assistant\n<think>\n' + reasoning_content + '\n</think>\n\n' + content + '<|im_end|>\n' -}}
+    {% endgeneration %}
+{%- endif -%}
+{# GENERATION PROMPT #}
+{%- if add_generation_prompt -%}
+    {{- '<|im_start|>assistant\n' -}}
+    {%- if not enable_thinking -%}
+        {{- '<think>\n\n</think>\n\n' -}}
+    {%- else %}
+        {{- '<think>\n' -}}
+    {%- endif %}
+{%- endif -%}

config.json

@@ -0,0 +1,122 @@
+{
+    "architectures": [
+        "Qwen3_5ForConditionalGeneration"
+    ],
+    "bos_token_id": null,
+    "dtype": "bfloat16",
+    "eos_token_id": 248044,
+    "image_token_id": 248056,
+    "model_type": "qwen3_5",
+    "pad_token_id": 248044,
+    "quantization": {
+        "group_size": 64,
+        "bits": 5,
+        "mode": "affine"
+    },
+    "quantization_config": {
+        "group_size": 64,
+        "bits": 5,
+        "mode": "affine"
+    },
+    "text_config": {
+        "attention_bias": false,
+        "attention_dropout": 0.0,
+        "attn_output_gate": true,
+        "bos_token_id": null,
+        "dtype": "bfloat16",
+        "eos_token_id": 248044,
+        "full_attention_interval": 4,
+        "head_dim": 256,
+        "hidden_act": "silu",
+        "hidden_size": 2560,
+        "initializer_range": 0.02,
+        "intermediate_size": 9216,
+        "layer_types": [
+            "linear_attention",
+            "linear_attention",
+            "linear_attention",
+            "full_attention",
+            "linear_attention",
+            "linear_attention",
+            "linear_attention",
+            "full_attention",
+            "linear_attention",
+            "linear_attention",
+            "linear_attention",
+            "full_attention",
+            "linear_attention",
+            "linear_attention",
+            "linear_attention",
+            "full_attention",
+            "linear_attention",
+            "linear_attention",
+            "linear_attention",
+            "full_attention",
+            "linear_attention",
+            "linear_attention",
+            "linear_attention",
+            "full_attention",
+            "linear_attention",
+            "linear_attention",
+            "linear_attention",
+            "full_attention",
+            "linear_attention",
+            "linear_attention",
+            "linear_attention",
+            "full_attention"
+        ],
+        "linear_conv_kernel_dim": 4,
+        "linear_key_head_dim": 128,
+        "linear_num_key_heads": 16,
+        "linear_num_value_heads": 32,
+        "linear_value_head_dim": 128,
+        "mamba_ssm_dtype": "float32",
+        "max_position_embeddings": 262144,
+        "mlp_only_layers": [],
+        "model_type": "qwen3_5_text",
+        "mtp_num_hidden_layers": 1,
+        "mtp_use_dedicated_embeddings": false,
+        "num_attention_heads": 16,
+        "num_hidden_layers": 32,
+        "num_key_value_heads": 4,
+        "pad_token_id": null,
+        "partial_rotary_factor": 0.25,
+        "rms_norm_eps": 1e-06,
+        "rope_parameters": {
+            "mrope_interleaved": true,
+            "mrope_section": [
+                11,
+                11,
+                10
+            ],
+            "partial_rotary_factor": 0.25,
+            "rope_theta": 10000000,
+            "rope_type": "default"
+        },
+        "tie_word_embeddings": true,
+        "use_cache": true,
+        "vocab_size": 248320
+    },
+    "tie_word_embeddings": true,
+    "transformers_version": "5.5.4",
+    "video_token_id": 248057,
+    "vision_config": {
+        "deepstack_visual_indexes": [],
+        "depth": 24,
+        "dtype": "bfloat16",
+        "hidden_act": "gelu_pytorch_tanh",
+        "hidden_size": 1024,
+        "in_channels": 3,
+        "initializer_range": 0.02,
+        "intermediate_size": 4096,
+        "model_type": "qwen3_5",
+        "num_heads": 16,
+        "num_position_embeddings": 2304,
+        "out_hidden_size": 2560,
+        "patch_size": 16,
+        "spatial_merge_size": 2,
+        "temporal_patch_size": 2
+    },
+    "vision_end_token_id": 248054,
+    "vision_start_token_id": 248053
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "eos_token_id": 248044,
+  "transformers_version": "5.5.4",
+  "use_cache": true
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:59d7a10a7a5ff6da03d3d31ffe418de847a6ad53328aaec5360fcebca3dc0465
+size 3559898117

processor_config.json

@@ -0,0 +1,50 @@
+{
+  "image_processor": {
+    "do_convert_rgb": true,
+    "do_normalize": true,
+    "do_rescale": true,
+    "image_mean": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "image_processor_type": "Qwen3VLImageProcessor",
+    "image_std": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "max_pixels": 16777216,
+    "merge_size": 2,
+    "min_pixels": 65536,
+    "patch_size": 16,
+    "rescale_factor": 0.00392156862745098,
+    "temporal_patch_size": 2
+  },
+  "processor_class": "Qwen3VLProcessor",
+  "video_processor": {
+    "do_convert_rgb": true,
+    "do_normalize": true,
+    "do_rescale": true,
+    "fps": 2.0,
+    "image_mean": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "image_std": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "max_frames": 768,
+    "max_pixels": 786432,
+    "merge_size": 2,
+    "min_frames": 4,
+    "min_pixels": 131072,
+    "patch_size": 16,
+    "rescale_factor": 0.00392156862745098,
+    "temporal_patch_size": 2,
+    "video_processor_type": "Qwen3VLVideoProcessor"
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:87a7830d63fcf43bf241c3c5242e96e62dd3fdc29224ca26fed8ea333db72de4
+size 19989343

tokenizer_config.json

@@ -0,0 +1,33 @@
+{
+  "add_prefix_space": false,
+  "audio_bos_token": "<|audio_start|>",
+  "audio_eos_token": "<|audio_end|>",
+  "audio_token": "<|audio_pad|>",
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "image_token": "<|image_pad|>",
+  "is_local": true,
+  "local_files_only": false,
+  "model_max_length": 262144,
+  "model_specific_special_tokens": {
+    "audio_bos_token": "<|audio_start|>",
+    "audio_eos_token": "<|audio_end|>",
+    "audio_token": "<|audio_pad|>",
+    "image_token": "<|image_pad|>",
+    "video_token": "<|video_pad|>",
+    "vision_bos_token": "<|vision_start|>",
+    "vision_eos_token": "<|vision_end|>"
+  },
+  "pad_token": "<|endoftext|>",
+  "pretokenize_regex": "(?i:'s|'t|'re|'ve|'m|'ll|'d)|[^\\r\\n\\p{L}\\p{N}]?[\\p{L}\\p{M}]+|\\p{N}| ?[^\\s\\p{L}\\p{M}\\p{N}]+[\\r\\n]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
+  "processor_class": "Qwen3VLProcessor",
+  "split_special_tokens": false,
+  "tokenizer_class": "TokenizersBackend",
+  "unk_token": null,
+  "video_token": "<|video_pad|>",
+  "vision_bos_token": "<|vision_start|>",
+  "vision_eos_token": "<|vision_end|>"
+}