Upload processor

chat_template.jinja

@@ -0,0 +1,81 @@
+
+{%- set image_count = namespace(value=0) -%}
+{%- set video_count = namespace(value=0) -%}
+{{- '<|begin_of_sentence|>' }}
+{%- for message in messages -%}
+    {%- if message.role in ['system', 'user'] -%}
+        {%- if message.role == 'user' -%}
+            {{- 'User: ' -}}
+        {%- endif -%}
+        {%- if message.content is string -%}
+            {{- message.content -}}
+        {%- else -%}
+            {%- for content_item in message.content -%}
+                {%- if content_item.type == 'text' -%}
+                    {{- content_item.text -}}
+                {%- elif content_item.type in ['image_url', 'image'] -%}
+                    {%- set image_count.value = image_count.value + 1 -%}
+                    Picture {{ image_count.value }}:<|IMAGE_START|><|IMAGE_PLACEHOLDER|><|IMAGE_END|>
+                {%- elif content_item.type in ['video_url', 'video'] -%}
+                    {%- set video_count.value = video_count.value + 1 -%}
+                    Video {{ video_count.value }}:<|VIDEO_START|><|VIDEO_PLACEHOLDER|><|VIDEO_END|>
+                {%- endif -%}
+            {%- endfor -%}
+        {%- endif -%}
+        {%- if message.role == 'system' -%}
+            {{- '
+            ' -}}
+        {%- endif -%}
+    {%- elif message.role == 'assistant' -%}
+        {%- macro extract_text_content(content_field) -%}
+            {%- if content_field is string -%}
+                {{- content_field -}}
+            {%- elif content_field is iterable and content_field is not string -%}
+                {%- set ns = namespace(text_parts=[]) -%}
+                {%- set text_parts = [] -%}
+                {%- for item in content_field -%}
+                    {%- if item.type == 'text' -%}
+                        {%- set ns.text_parts = ns.text_parts + [item.text] -%}
+                    {%- endif -%}
+                {%- endfor -%}
+                {{- ns.text_parts | join("") -}}
+            {%- else -%}
+                {{- '' -}}
+            {%- endif -%}
+        {%- endmacro -%}
+        {%- set reasoning_content = extract_text_content(message.reasoning_content) -%}
+        {%- set content = extract_text_content(message.content) -%}
+        {%- if '</think>' in content %}
+            {%- set reasoning_content = content.split('</think>')[0].rstrip('
+                        ').split('<think>')[-1].lstrip('
+                        ') %}
+            {%- set content = content.split('</think>')[-1].lstrip('
+                        ') %}
+        {%- endif %}
+        {%- if reasoning_content %}
+            {{- '
+            ' + 'Assistant: ' + '<think>
+            ' + reasoning_content.strip('
+                        ') + '
+            </think>
+            ' + content.lstrip('
+            ') }}
+        {%- else %}
+            {{- '
+            ' + 'Assistant: ' + content }}
+        {%- endif %}
+        {{- '<|end_of_sentence |>' }}
+    {%- endif -%}
+{%- endfor -%}
+{%- if add_generation_prompt is not defined or add_generation_prompt is true %}
+    {{- '
+Assistant: ' -}}
+    {%- if (enable_thinking is defined and enable_thinking is false) or enable_thinking is not defined %}
+        {{- '<think>
+
+</think>
+
+' }}
+    {%- endif %}
+    {%- if enable_thinking is defined and enable_thinking is true %}{{- '<think>' }}{%- endif %}
+{%- endif %}

processor_config.json

@@ -0,0 +1,65 @@
+{
+  "image_processor": {
+    "data_format": "channels_first",
+    "do_convert_rgb": true,
+    "do_normalize": true,
+    "do_rescale": true,
+    "do_resize": true,
+    "image_mean": [
+      0.48145466,
+      0.4578275,
+      0.40821073
+    ],
+    "image_processor_type": "Ernie4_5_VL_MoeImageProcessorFast",
+    "image_std": [
+      0.26862954,
+      0.26130258,
+      0.27577711
+    ],
+    "merge_size": 2,
+    "patch_size": 14,
+    "processor_class": "Ernie4_5_VL_MoeProcessor",
+    "resample": 3,
+    "rescale_factor": 0.00392156862745098,
+    "size": {
+      "longest_edge": 4842768,
+      "shortest_edge": 3136
+    }
+  },
+  "processor_class": "Ernie4_5_VL_MoeProcessor",
+  "video_processor": {
+    "data_format": "channels_first",
+    "default_to_square": true,
+    "do_convert_rgb": true,
+    "do_normalize": true,
+    "do_rescale": true,
+    "do_resize": true,
+    "do_sample_frames": true,
+    "draw_on_frames": true,
+    "font": "Roboto-Regular.ttf",
+    "image_mean": [
+      0.48145466,
+      0.4578275,
+      0.40821073
+    ],
+    "image_std": [
+      0.26862954,
+      0.26130258,
+      0.27577711
+    ],
+    "max_frames": 180,
+    "merge_size": 2,
+    "min_frames": 16,
+    "patch_size": 14,
+    "processor_class": "Ernie4_5_VL_MoeProcessor",
+    "resample": 3,
+    "rescale_factor": 0.00392156862745098,
+    "return_metadata": false,
+    "size": {
+      "longest_edge": 937664,
+      "shortest_edge": 234416
+    },
+    "temporal_patch_size": 2,
+    "video_processor_type": "Ernie4_5_VL_MoeVideoProcessor"
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:94718bd0ed3c1f3591a1cf2c517d6364283457f2fd7e95ae4d201138e922079d
+size 11590466

tokenizer_config.json

@@ -0,0 +1,49 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "cls_token": "<|begin_of_sentence|>",
+  "eos_token": "</s>",
+  "extra_special_tokens": [
+    "<|IMAGE_PLACEHOLDER|>",
+    "<|IMAGE_START|>",
+    "<|IMAGE_END|>",
+    "<|VIDEO_PLACEHOLDER|>",
+    "<|VIDEO_START|>",
+    "<|VIDEO_END|>",
+    "<think>",
+    "</think>"
+  ],
+  "header_end_token": "<mask:7>",
+  "header_start_token": "<mask:6>",
+  "image_end_token": "<|IMAGE_END|>",
+  "image_start_token": "<|IMAGE_START|>",
+  "image_token": "<|IMAGE_PLACEHOLDER|>",
+  "is_local": true,
+  "mask_token": "<mask:1>",
+  "model_max_length": 131072,
+  "model_specific_special_tokens": {
+    "header_end_token": "<mask:7>",
+    "header_start_token": "<mask:6>",
+    "image_end_token": "<|IMAGE_END|>",
+    "image_start_token": "<|IMAGE_START|>",
+    "image_token": "<|IMAGE_PLACEHOLDER|>",
+    "sys_end_token": "<mask:5>",
+    "sys_start_token": "<mask:4>",
+    "video_end_token": "<|VIDEO_END|>",
+    "video_start_token": "<|VIDEO_START|>",
+    "video_token": "<|VIDEO_PLACEHOLDER|>"
+  },
+  "pad_token": "<unk>",
+  "processor_class": "Ernie4_5_VL_MoeProcessor",
+  "sep_token": "<|end_of_sentence|>",
+  "sys_end_token": "<mask:5>",
+  "sys_start_token": "<mask:4>",
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": false,
+  "video_end_token": "<|VIDEO_END|>",
+  "video_start_token": "<|VIDEO_START|>",
+  "video_token": "<|VIDEO_PLACEHOLDER|>"
+}