Add tiny processor fixture for testing

.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

chat_template.jinja

@@ -0,0 +1,77 @@
+{%- if messages[0]['role'] == 'system' %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0]['content'] is string %}
+        {{- messages[0]['content'] }}
+    {%- else %}
+        {%- for item in messages[0]['content'] %}
+            {%- if item['type'] == 'text' %}
+                {{- item['text'] }}
+            {%- endif %}
+        {%- endfor %}
+    {%- endif %}
+    {{- '<|im_end|>\n' }}
+{%- else %}
+    {{- '<|im_start|>system\n你是Qianfan-VL，由百度智能云研发的多模态大语言模型。<|im_end|>\n' }}
+{%- endif %}
+{%- set ns = namespace(found_last_user=false, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+    {%- set index = (messages|length - 1) - loop.index0 %}
+    {%- if not ns.found_last_user and message['role'] == 'user' %}
+        {%- set ns.found_last_user = true %}
+        {%- set ns.last_query_index = index %}
+    {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+    {%- if messages[0]['role'] != 'system' or not loop.first %}
+        {%- if message['role'] == 'user' or (message['role'] == 'system' and not loop.first) %}
+            {%- set append_think = (enable_thinking is defined and enable_thinking and message['role'] == 'user' and loop.index0 == ns.last_query_index) %}
+                {{- '<|im_start|>' + message['role'] + '\n' }}
+            {%- if message['content'] is string %}
+                {{- message['content'] }}
+            {%- else %}
+            {%- for item in message['content'] %}
+                {%- if item['type'] == 'image' %}
+                    {{- '<image>\n' }}
+                {%- elif item['type'] == 'video' %}
+                    {{- '<video>\n' }}
+                {%- elif item['type'] == 'text' %}
+                    {{- item['text'] }}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+        {%- if append_think %}
+            {{- '<think>' }}
+        {%- endif %}
+        {{- '<|im_end|>\n' }}
+        {%- elif message['role'] == 'assistant' %}
+            {%- if message['content'] is string %}
+                {%- set raw_content = message['content'] %}
+            {%- else %}
+                {%- set content_ns = namespace(raw='') %}
+                {%- for item in message['content'] %}
+                    {%- if item['type'] == 'text' %}
+                        {%- set content_ns.raw = content_ns.raw + item['text'] %}
+                    {%- endif %}
+                {%- endfor %}
+                {%- set raw_content = content_ns.raw %}
+            {%- endif %}
+            {%- set content = raw_content %}
+            {%- set reasoning_content = '' %}
+            {%- if 'reasoning_content' in message and message['reasoning_content'] is not none %}
+                {%- set reasoning_content = message['reasoning_content'] %}
+            {%- elif '</think>' in raw_content %}
+                {%- set content = raw_content.split('</think>')[-1].lstrip('\n') %}
+                {%- set reasoning_content = raw_content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
+            {%- endif %}
+            {%- if loop.index0 > ns.last_query_index and reasoning_content %}
+                {{- '<|im_start|>' + message['role'] + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
+            {%- else %}
+                {{- '<|im_start|>' + message['role'] + '\n' + content }}
+            {%- endif %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+{%- endif %}

config.json

@@ -0,0 +1,78 @@
+{
+  "architectures": [
+    "QianfanOCRForConditionalGeneration"
+  ],
+  "downsample_ratio": 0.5,
+  "dynamic_image_size": true,
+  "eos_token_id": 502,
+  "force_image_size": 448,
+  "image_token_id": 151671,
+  "text_config": {
+    "_name_or_path": "/root/codespace/checkpoints/Qwen3-4B",
+    "architectures": [
+      "Qwen3ForCausalLM"
+    ],
+    "attention_bias": false,
+    "attention_dropout": 0.0,
+    "bos_token_id": 151643,
+    "debug": false,
+    "eos_token_id": 151645,
+    "ep_size": 1,
+    "head_dim": 128,
+    "hidden_act": "silu",
+    "hidden_size": 2560,
+    "initializer_range": 0.02,
+    "intermediate_size": 9728,
+    "max_position_embeddings": 32768,
+    "max_window_layers": 36,
+    "micro_forward": false,
+    "model_type": "qwen3",
+    "num_attention_heads": 32,
+    "num_hidden_layers": 36,
+    "num_key_value_heads": 8,
+    "rms_norm_eps": 1e-06,
+    "rope_scaling": null,
+    "rope_theta": 5000000,
+    "skip_checkpoint": false,
+    "sliding_window": null,
+    "torch_dtype": "bfloat16",
+    "use_cache": false,
+    "use_deepep": false,
+    "use_sliding_window": false,
+    "vocab_size": 153678
+  },
+  "max_dynamic_patch": 12,
+  "min_dynamic_patch": 1,
+  "model_type": "qianfan_ocr",
+  "pad2square": false,
+  "pad_token_id": 500,
+  "ps_version": "v2",
+  "select_layer": -1,
+  "template": "internvl2_5",
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": null,
+  "use_thumbnail": true,
+  "vision_config": {
+    "architectures": [
+      "QianfanOCRVisionModel"
+    ],
+    "attention_dropout": 0.0,
+    "drop_path_rate": 0.1,
+    "hidden_act": "gelu",
+    "hidden_size": 1024,
+    "image_size": 448,
+    "initializer_range": 0.02,
+    "intermediate_size": 4096,
+    "layer_norm_eps": 1e-06,
+    "model_type": "qianfan_ocr_vision",
+    "norm_type": "layer_norm",
+    "num_attention_heads": 16,
+    "num_channels": 3,
+    "num_hidden_layers": 24,
+    "patch_size": 14,
+    "qkv_bias": true,
+    "torch_dtype": "bfloat16"
+  },
+  "vocab_size": 500
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "transformers_version": "4.55.0"
+}

preprocessor_config.json

@@ -0,0 +1,34 @@
+{
+  "crop_size": null,
+  "crop_to_patches": false,
+  "data_format": "channels_first",
+  "default_to_square": true,
+  "device": null,
+  "do_center_crop": null,
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.485,
+    0.456,
+    0.406
+  ],
+  "image_processor_type": "GotOcr2ImageProcessor",
+  "image_std": [
+    0.229,
+    0.224,
+    0.225
+  ],
+  "input_data_format": null,
+  "max_patches": 12,
+  "min_patches": 1,
+  "processor_class": "InternVLProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "return_tensors": null,
+  "size": {
+    "height": 448,
+    "width": 448
+  }
+}

processor_config.json

@@ -0,0 +1,7 @@
+{
+  "image_seq_length": 256,
+  "processor_class": "QianfanOCRProcessor",
+  "start_image_token": "<img>",
+  "end_image_token": "</img>",
+  "context_image_token": "<IMG_CONTEXT>"
+}

video_preprocessor_config.json

@@ -0,0 +1,70 @@
+{
+  "_valid_kwargs_names": [
+    "do_convert_rgb",
+    "do_resize",
+    "size",
+    "size_divisor",
+    "default_to_square",
+    "resample",
+    "do_rescale",
+    "rescale_factor",
+    "do_normalize",
+    "image_mean",
+    "image_std",
+    "do_pad",
+    "do_center_crop",
+    "crop_size",
+    "data_format",
+    "input_data_format",
+    "device"
+  ],
+  "crop_size": null,
+  "data_format": "channels_first",
+  "default_to_square": true,
+  "device": null,
+  "do_center_crop": null,
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_pad": null,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "input_data_format": null,
+  "model_valid_processing_keys": [
+    "do_convert_rgb",
+    "do_resize",
+    "size",
+    "size_divisor",
+    "default_to_square",
+    "resample",
+    "do_rescale",
+    "rescale_factor",
+    "do_normalize",
+    "image_mean",
+    "image_std",
+    "do_pad",
+    "do_center_crop",
+    "crop_size",
+    "data_format",
+    "input_data_format",
+    "device"
+  ],
+  "processor_class": "InternVLProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "height": 384,
+    "width": 384
+  },
+  "size_divisor": null,
+  "video_processor_type": "InternVLVideoProcessor"
+}