Upload folder using huggingface_hub

.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 +1,193 @@
+---
+library_name: transformers
+pipeline_tag: image-text-to-text
+inference: true
+widget:
+  - text: Hello!
+    example_title: Hello world
+    group: Python
+base_model:
+- THUDM/GLM-4.1V-9B-Thinking
+---
+
+This tiny model is for debugging. It is randomly initialized with the config adapted from [THUDM/GLM-4.1V-9B-Thinking](https://huggingface.co/THUDM/GLM-4.1V-9B-Thinking).
+
+### Example usage:
+
+```python
+import os
+import re
+
+import torch
+
+from transformers import AutoProcessor, Glm4vForConditionalGeneration
+
+model_id = "tiny-random/glm-4v"
+messages = [
+    {
+        "role": "user",
+        "content": [
+            {
+                "type": "image",
+                "url": "https://upload.wikimedia.org/wikipedia/commons/f/fa/Grayscale_8bits_palette_sample_image.png"
+            },
+            {
+                "type": "text",
+                "text": "describe this image"
+            }
+        ],
+    }
+]
+processor = AutoProcessor.from_pretrained(model_id)
+model = Glm4vForConditionalGeneration.from_pretrained(
+    pretrained_model_name_or_path=model_id,
+    torch_dtype=torch.bfloat16,
+    device_map="auto",
+)
+inputs = processor.apply_chat_template(
+    messages,
+    tokenize=True,
+    add_generation_prompt=True,
+    return_dict=True,
+    return_tensors="pt"
+).to(model.device)
+generated_ids = model.generate(**inputs, max_new_tokens=16)
+output_text = processor.decode(generated_ids[0][inputs["input_ids"].shape[1]:], skip_special_tokens=False)
+print(output_text)
+```
+
+### Codes to create this repo:
+
+```python
+import json
+from pathlib import Path
+
+import torch
+
+import accelerate
+from huggingface_hub import file_exists, hf_hub_download
+from transformers import (
+    AutoConfig,
+    AutoModelForCausalLM,
+    AutoProcessor,
+    GenerationConfig,
+    set_seed,
+)
+from transformers import AutoProcessor, Glm4vForConditionalGeneration
+
+source_model_id = "THUDM/GLM-4.1V-9B-Thinking"
+save_folder = "/tmp/tiny-random/glm-4v"
+
+processor = AutoProcessor.from_pretrained(source_model_id, trust_remote_code=True)
+processor.save_pretrained(save_folder)
+
+with open(hf_hub_download(source_model_id, filename='config.json', repo_type='model'), 'r', encoding='utf-8') as f:
+    config_json = json.load(f)
+config_json['hidden_size'] = 64
+config_json['intermediate_size'] = 128
+config_json['num_attention_heads'] = 2
+config_json['num_hidden_layers'] = 2
+config_json['num_key_value_heads'] = 1
+config_json['tie_word_embeddings'] = True
+config_json['vision_config']['hidden_size'] = 64
+config_json['vision_config']['depth'] = 2
+config_json['vision_config']['num_heads'] = 2
+config_json['vision_config']['intermediate_size'] = 128
+config_json['vision_config']['out_hidden_size'] = 64
+config_json['rope_scaling']['mrope_section'] = [2, 2, 4]
+
+with open(f"{save_folder}/config.json", "w", encoding='utf-8') as f:
+    json.dump(config_json, f, indent=2)
+
+config = AutoConfig.from_pretrained(
+    save_folder,
+    trust_remote_code=True,
+)
+print(config)
+torch.set_default_dtype(torch.bfloat16)
+model = Glm4vForConditionalGeneration(config)
+torch.set_default_dtype(torch.float32)
+if file_exists(filename="generation_config.json", repo_id=source_model_id, repo_type='model'):
+    model.generation_config = GenerationConfig.from_pretrained(
+        source_model_id, trust_remote_code=True,
+    )
+set_seed(42)
+model = model.cpu()  # cpu is more stable for random initialization across machines
+with torch.no_grad():
+    for name, p in sorted(model.named_parameters()):
+        torch.nn.init.normal_(p, 0, 0.2)
+        print(name, p.shape)
+model.save_pretrained(save_folder)
+print(model)
+```
+
+### Printing the model:
+
+```text
+Glm4vForConditionalGeneration(
+  (model): Glm4vModel(
+    (visual): Glm4vVisionModel(
+      (embeddings): Glm4vVisionEmbeddings(
+        (position_embedding): Embedding(576, 64)
+      )
+      (patch_embed): Glm4vVisionPatchEmbed(
+        (proj): Conv3d(3, 64, kernel_size=(2, 14, 14), stride=(2, 14, 14))
+      )
+      (rotary_pos_emb): Glm4vVisionRotaryEmbedding()
+      (blocks): ModuleList(
+        (0-1): 2 x Glm4vVisionBlock(
+          (norm1): Glm4vRMSNorm((64,), eps=1e-05)
+          (norm2): Glm4vRMSNorm((64,), eps=1e-05)
+          (attn): Glm4vVisionAttention(
+            (qkv): Linear(in_features=64, out_features=192, bias=False)
+            (proj): Linear(in_features=64, out_features=64, bias=False)
+          )
+          (mlp): Glm4VisionMlp(
+            (gate_proj): Linear(in_features=64, out_features=64, bias=False)
+            (up_proj): Linear(in_features=64, out_features=64, bias=False)
+            (down_proj): Linear(in_features=64, out_features=64, bias=False)
+            (act_fn): SiLU()
+          )
+        )
+      )
+      (merger): Glm4vVisionPatchMerger(
+        (proj): Linear(in_features=64, out_features=64, bias=False)
+        (post_projection_norm): LayerNorm((64,), eps=1e-05, elementwise_affine=True)
+        (gate_proj): Linear(in_features=64, out_features=128, bias=False)
+        (up_proj): Linear(in_features=64, out_features=128, bias=False)
+        (down_proj): Linear(in_features=128, out_features=64, bias=False)
+        (act1): GELU(approximate='none')
+        (act_fn): SiLU()
+      )
+      (post_conv_layernorm): Glm4vRMSNorm((64,), eps=1e-05)
+      (downsample): Conv2d(64, 64, kernel_size=(2, 2), stride=(2, 2))
+      (post_layernorm): Glm4vRMSNorm((64,), eps=1e-05)
+    )
+    (language_model): Glm4vTextModel(
+      (embed_tokens): Embedding(151552, 64, padding_idx=151329)
+      (layers): ModuleList(
+        (0-1): 2 x Glm4vTextDecoderLayer(
+          (self_attn): Glm4vTextAttention(
+            (q_proj): Linear(in_features=64, out_features=64, bias=True)
+            (k_proj): Linear(in_features=64, out_features=32, bias=True)
+            (v_proj): Linear(in_features=64, out_features=32, bias=True)
+            (o_proj): Linear(in_features=64, out_features=64, bias=False)
+          )
+          (mlp): Glm4vTextMLP(
+            (gate_up_proj): Linear(in_features=64, out_features=256, bias=False)
+            (down_proj): Linear(in_features=128, out_features=64, bias=False)
+            (activation_fn): SiLU()
+          )
+          (input_layernorm): Glm4vRMSNorm((64,), eps=1e-05)
+          (post_attention_layernorm): Glm4vRMSNorm((64,), eps=1e-05)
+          (post_self_attn_layernorm): Glm4vRMSNorm((64,), eps=1e-05)
+          (post_mlp_layernorm): Glm4vRMSNorm((64,), eps=1e-05)
+        )
+      )
+      (norm): Glm4vRMSNorm((64,), eps=1e-05)
+      (rotary_emb): Glm4vTextRotaryEmbedding()
+    )
+  )
+  (lm_head): Linear(in_features=64, out_features=151552, bias=False)
+)
+```

chat_template.jinja

@@ -0,0 +1,33 @@
+[gMASK]<sop>
+{%- for msg in messages %}
+    {%- if msg.role == 'system' %}
+<|system|>
+{{ msg.content }}
+    {%- elif msg.role == 'user' %}
+<|user|>{{ '\n' }}
+
+        {%- if msg.content is string %}
+{{ msg.content }}
+        {%- else %}
+            {%- for item in msg.content %}
+                {%- if item.type == 'video' or 'video' in item %}
+<|begin_of_video|><|video|><|end_of_video|>
+                {%- elif item.type == 'image' or 'image' in item %}
+<|begin_of_image|><|image|><|end_of_image|>
+                {%- elif item.type == 'text' %}
+{{ item.text }}
+                {%- endif %}
+            {%- endfor %}
+        {%- endif %}
+    {%- elif msg.role == 'assistant' %}
+        {%- if msg.metadata %}
+<|assistant|>{{ msg.metadata }}
+{{ msg.content }}
+        {%- else %}
+<|assistant|>
+{{ msg.content }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{% if add_generation_prompt %}<|assistant|>
+{% endif %}

config.json

@@ -0,0 +1,105 @@
+{
+  "architectures": [
+    "Glm4vForConditionalGeneration"
+  ],
+  "attention_bias": true,
+  "attention_dropout": 0.0,
+  "eos_token_id": [
+    151329,
+    151336,
+    151338,
+    151348
+  ],
+  "hidden_act": "silu",
+  "hidden_size": 64,
+  "image_end_token_id": 151340,
+  "image_start_token_id": 151339,
+  "image_token_id": 151343,
+  "initializer_range": 0.02,
+  "intermediate_size": 128,
+  "max_position_embeddings": 65536,
+  "model_type": "glm4v",
+  "num_attention_heads": 2,
+  "num_hidden_layers": 2,
+  "num_key_value_heads": 1,
+  "pad_token_id": 151329,
+  "partial_rotary_factor": 0.5,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "mrope_section": [
+      2,
+      2,
+      4
+    ],
+    "rope_type": "default",
+    "type": "default"
+  },
+  "rope_theta": 10000.0,
+  "text_config": {
+    "architectures": [
+      "Glm4vForConditionalGeneration"
+    ],
+    "attention_bias": true,
+    "attention_dropout": 0.0,
+    "eos_token_id": [
+      151329,
+      151336,
+      151338,
+      151348
+    ],
+    "hidden_act": "silu",
+    "hidden_size": 64,
+    "image_token_id": null,
+    "initializer_range": 0.02,
+    "intermediate_size": 128,
+    "max_position_embeddings": 65536,
+    "model_type": "glm4v_text",
+    "num_attention_heads": 2,
+    "num_hidden_layers": 2,
+    "num_key_value_heads": 1,
+    "pad_token_id": 151329,
+    "partial_rotary_factor": 0.5,
+    "rms_norm_eps": 1e-05,
+    "rope_scaling": {
+      "mrope_section": [
+        2,
+        2,
+        4
+      ],
+      "rope_type": "default",
+      "type": "default"
+    },
+    "rope_theta": 10000.0,
+    "tie_word_embeddings": true,
+    "torch_dtype": "bfloat16",
+    "use_cache": true,
+    "video_token_id": null,
+    "vocab_size": 151552
+  },
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.54.0.dev0",
+  "use_cache": true,
+  "video_end_token_id": 151342,
+  "video_start_token_id": 151341,
+  "video_token_id": 151344,
+  "vision_config": {
+    "attention_bias": false,
+    "attention_dropout": 0.0,
+    "depth": 2,
+    "hidden_act": "silu",
+    "hidden_dropout_prob": 0.0,
+    "hidden_size": 64,
+    "image_size": 336,
+    "in_channels": 3,
+    "initializer_range": 0.02,
+    "intermediate_size": 128,
+    "model_type": "glm4v",
+    "num_heads": 2,
+    "out_hidden_size": 64,
+    "patch_size": 14,
+    "rms_norm_eps": 1e-05,
+    "spatial_merge_size": 2,
+    "temporal_patch_size": 2
+  },
+  "vocab_size": 151552
+}

generation_config.json

@@ -0,0 +1,11 @@
+{
+  "_from_model_config": true,
+  "eos_token_id": [
+    151329,
+    151336,
+    151338,
+    151348
+  ],
+  "pad_token_id": 151329,
+  "transformers_version": "4.54.0.dev0"
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a51b22dc53e1de592646c2a898839abada0c547804444a118dac2508ec322d5c
+size 19984480

preprocessor_config.json

@@ -0,0 +1,27 @@
+{
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_processor_type": "Glm4vImageProcessor",
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "merge_size": 2,
+  "patch_size": 14,
+  "processor_class": "Glm4vProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "longest_edge": 11760000,
+    "shortest_edge": 12544
+  },
+  "temporal_patch_size": 2
+}

special_tokens_map.json

@@ -0,0 +1,34 @@
+{
+  "additional_special_tokens": [
+    "<|endoftext|>",
+    "[MASK]",
+    "[gMASK]",
+    "[sMASK]",
+    "<sop>",
+    "<eop>",
+    "<|system|>",
+    "<|user|>",
+    "<|assistant|>",
+    "<|observation|>",
+    "<|begin_of_image|>",
+    "<|end_of_image|>",
+    "<|begin_of_video|>",
+    "<|end_of_video|>",
+    "<|image|>",
+    "<|video|>"
+  ],
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bcc742fc44db1f0870d7320b495e6240cc40e202565fc96786220fa0d9ddb41c
+size 19968183

tokenizer_config.json

@@ -0,0 +1,220 @@
+{
+  "added_tokens_decoder": {
+    "151329": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151330": {
+      "content": "[MASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151331": {
+      "content": "[gMASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151332": {
+      "content": "[sMASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151333": {
+      "content": "<sop>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151334": {
+      "content": "<eop>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151335": {
+      "content": "<|system|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151336": {
+      "content": "<|user|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151337": {
+      "content": "<|assistant|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151338": {
+      "content": "<|observation|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151339": {
+      "content": "<|begin_of_image|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151340": {
+      "content": "<|end_of_image|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151341": {
+      "content": "<|begin_of_video|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151342": {
+      "content": "<|end_of_video|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151343": {
+      "content": "<|image|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151344": {
+      "content": "<|video|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151345": {
+      "content": "<think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151346": {
+      "content": "</think>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151347": {
+      "content": "<answer>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151348": {
+      "content": "</answer>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151349": {
+      "content": "<|begin_of_box|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151350": {
+      "content": "<|end_of_box|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151351": {
+      "content": "<|sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|endoftext|>",
+    "[MASK]",
+    "[gMASK]",
+    "[sMASK]",
+    "<sop>",
+    "<eop>",
+    "<|system|>",
+    "<|user|>",
+    "<|assistant|>",
+    "<|observation|>",
+    "<|begin_of_image|>",
+    "<|end_of_image|>",
+    "<|begin_of_video|>",
+    "<|end_of_video|>",
+    "<|image|>",
+    "<|video|>"
+  ],
+  "clean_up_tokenization_spaces": false,
+  "do_lower_case": false,
+  "eos_token": "<|endoftext|>",
+  "extra_special_tokens": {},
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 65536,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "left",
+  "processor_class": "Glm4vProcessor",
+  "remove_space": false,
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}

video_preprocessor_config.json

@@ -0,0 +1,42 @@
+{
+  "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,
+  "do_sample_frames": true,
+  "fps": 2,
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "input_data_format": null,
+  "max_image_size": {
+    "longest_edge": 47040000
+  },
+  "merge_size": 2,
+  "num_frames": 16,
+  "patch_size": 14,
+  "processor_class": "Glm4vProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "longest_edge": 47040000,
+    "shortest_edge": 12544
+  },
+  "size_divisor": null,
+  "temporal_patch_size": 2,
+  "video_metadata": null,
+  "video_processor_type": "Glm4vVideoProcessor"
+}