Upload folder using huggingface_hub

config.json

@@ -16,20 +16,19 @@
   "image_start_token_id": 151339,
   "image_token_id": 151343,
   "initializer_range": 0.02,
-  "intermediate_size": 128,
+  "intermediate_size": 64,
   "max_position_embeddings": 65536,
   "model_type": "glm4v",
-  "num_attention_heads": 2,
-  "num_hidden_layers": 2,
-  "num_key_value_heads": 1,
+  "num_attention_heads": 8,
+  "num_hidden_layers": 3,
+  "num_key_value_heads": 2,
   "pad_token_id": 151329,
   "partial_rotary_factor": 0.5,
   "rms_norm_eps": 1e-05,
   "rope_scaling": {
     "mrope_section": [
-      2,
-      2,
-      4
+      1,
+      1
     ],
     "rope_type": "default",
     "type": "default"
@@ -51,33 +50,32 @@
     "hidden_size": 64,
     "image_token_id": null,
     "initializer_range": 0.02,
-    "intermediate_size": 128,
+    "intermediate_size": 64,
     "max_position_embeddings": 65536,
     "model_type": "glm4v_text",
-    "num_attention_heads": 2,
-    "num_hidden_layers": 2,
-    "num_key_value_heads": 1,
+    "num_attention_heads": 8,
+    "num_hidden_layers": 3,
+    "num_key_value_heads": 2,
     "pad_token_id": 151329,
     "partial_rotary_factor": 0.5,
     "rms_norm_eps": 1e-05,
     "rope_scaling": {
       "mrope_section": [
-        2,
-        2,
-        4
+        1,
+        1
       ],
       "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",
+  "tie_word_embeddings": false,
+  "torch_dtype": "float32",
+  "transformers_version": "4.54.1",
   "use_cache": true,
   "video_end_token_id": 151342,
   "video_start_token_id": 151341,
@@ -85,17 +83,17 @@
   "vision_config": {
     "attention_bias": false,
     "attention_dropout": 0.0,
-    "depth": 2,
+    "depth": 4,
     "hidden_act": "silu",
     "hidden_dropout_prob": 0.0,
     "hidden_size": 64,
     "image_size": 336,
     "in_channels": 3,
     "initializer_range": 0.02,
-    "intermediate_size": 128,
+    "intermediate_size": 64,
     "model_type": "glm4v",
-    "num_heads": 2,
-    "out_hidden_size": 64,
+    "num_heads": 4,
+    "out_hidden_size": 4096,
     "patch_size": 14,
     "rms_norm_eps": 1e-05,
     "spatial_merge_size": 2,

generation_config.json

@@ -7,5 +7,5 @@
     151348
   ],
   "pad_token_id": 151329,
-  "transformers_version": "4.54.0.dev0"
+  "transformers_version": "4.54.1"
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a51b22dc53e1de592646c2a898839abada0c547804444a118dac2508ec322d5c
-size 19984480
+oid sha256:91d6edb23f856c096ba6962b6c7ea1d8d8336e16ef4dee2bcb8ae08733b66ea9
+size 146137472

preprocessor_config.json

@@ -1,27 +1,11 @@
 {
-  "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
-}
+    "size": {"shortest_edge": 12544, "longest_edge": 9633792},
+    "do_rescale": true,
+    "patch_size": 14,
+    "temporal_patch_size": 2,
+    "merge_size": 2,
+    "image_mean": [0.48145466, 0.4578275, 0.40821073],
+    "image_std": [0.26862954, 0.26130258, 0.27577711],
+    "image_processor_type": "Glm4vImageProcessor",
+    "processor_class": "Glm4vProcessor"
+}

tokenizer.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bcc742fc44db1f0870d7320b495e6240cc40e202565fc96786220fa0d9ddb41c
-size 19968183
+oid sha256:76ebeac0d8bd7879ead7b43c16b44981f277e47225de2bd7de9ae1a6cc664a8c
+size 19966496

tokenizer_config.json

@@ -112,7 +112,7 @@
       "single_word": false,
       "special": true
     },
-    "151343": {
+     "151343": {
       "content": "<|image|>",
       "lstrip": false,
       "normalized": false,
@@ -128,7 +128,7 @@
       "single_word": false,
       "special": true
     },
-    "151345": {
+     "151345": {
       "content": "<think>",
       "lstrip": false,
       "normalized": false,
@@ -136,7 +136,7 @@
       "single_word": false,
       "special": false
     },
-    "151346": {
+     "151346": {
       "content": "</think>",
       "lstrip": false,
       "normalized": false,
@@ -144,7 +144,7 @@
       "single_word": false,
       "special": false
     },
-    "151347": {
+     "151347": {
       "content": "<answer>",
       "lstrip": false,
       "normalized": false,
@@ -152,7 +152,7 @@
       "single_word": false,
       "special": false
     },
-    "151348": {
+     "151348": {
       "content": "</answer>",
       "lstrip": false,
       "normalized": false,
@@ -206,15 +206,13 @@
   "clean_up_tokenization_spaces": false,
   "do_lower_case": false,
   "eos_token": "<|endoftext|>",
-  "extra_special_tokens": {},
+  "pad_token": "<|endoftext|>",
   "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"
+  "tokenizer_class": "PreTrainedTokenizer"
 }

video_preprocessor_config.json

@@ -1,42 +1,11 @@
 {
-  "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"
-}
+    "size": {"shortest_edge": 12544, "longest_edge": 47040000},
+    "do_rescale": true,
+    "patch_size": 14,
+    "temporal_patch_size": 2,
+    "merge_size": 2,
+    "image_mean": [0.48145466, 0.4578275, 0.40821073],
+    "image_std": [0.26862954, 0.26130258, 0.27577711],
+    "video_processor_type": "Glm4vVideoProcessor",
+    "processor_class": "Glm4vProcessor"
+}