Upload folder using huggingface_hub

stage1-InternVideo2-1B-Qwen2_5-7B-tome16_mlp/config.json

@@ -0,0 +1,76 @@
+{
+  "_name_or_path": "/mnt/petrelfs/share_data/lixinhao/models/Qwen/Qwen2.5-7B-Instruct",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "frame_aspect_ratio": "square",
+  "frame_grid_pinpoints": null,
+  "hidden_act": "silu",
+  "hidden_size": 3584,
+  "image_aspect_ratio": "square",
+  "image_crop_resolution": null,
+  "image_grid_pinpoints": null,
+  "image_split_resolution": null,
+  "initializer_range": 0.02,
+  "intermediate_size": 18944,
+  "llm_compress_layer_list": [
+    8,
+    16,
+    24
+  ],
+  "llm_compress_type": "attention",
+  "llm_image_token_ratio_list": [
+    1.0,
+    0.5,
+    0.25,
+    0.125
+  ],
+  "max_num_pixels": 14745600000,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 28,
+  "min_slow_num_frames": 4,
+  "mm_close_init": false,
+  "mm_hidden_size": 1408,
+  "mm_llm_compress": false,
+  "mm_local_num_frames": 4,
+  "mm_newline_position": "one_token",
+  "mm_num_compress_latents": 128,
+  "mm_num_compress_query_type": "learnable",
+  "mm_patch_merge_type": "flat",
+  "mm_pos_num_frames": 4,
+  "mm_projector_lr": null,
+  "mm_projector_type": "tome16_mlp_hd64",
+  "mm_resampler_type": null,
+  "mm_spatial_pool_mode": "bilinear",
+  "mm_tunable_parts": "mm_mlp_adapter",
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -2,
+  "mm_vision_tower": "internvideo2",
+  "mm_vision_tower_lr": null,
+  "model_type": "qwen2",
+  "num_attention_heads": 28,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 4,
+  "pos_skipping_range": 4096,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": 131072,
+  "tie_word_embeddings": false,
+  "tokenizer_model_max_length": 8192,
+  "tokenizer_padding_side": "right",
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.39.2",
+  "use_cache": true,
+  "use_mm_proj": true,
+  "use_pos_skipping": false,
+  "use_sliding_window": false,
+  "vision_encode_type": "video_image",
+  "vision_tower_pretrained": null,
+  "vocab_size": 152064
+}

stage1-InternVideo2-1B-Qwen2_5-7B-tome16_mlp/mm_projector.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0df2f576b955638cb59f41b822f20a71dcdadcf14875bc2b6caf6ac7b66c1903
+size 35799096

stage1-UMT-Qwen2_5_2B-tome16_mlp/config.json

@@ -0,0 +1,76 @@
+{
+  "_name_or_path": "/mnt/petrelfs/share_data/lixinhao/models/Qwen/Qwen2.5-1.5B-Instruct",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "frame_aspect_ratio": "square",
+  "frame_grid_pinpoints": null,
+  "hidden_act": "silu",
+  "hidden_size": 1536,
+  "image_aspect_ratio": "square",
+  "image_crop_resolution": null,
+  "image_grid_pinpoints": null,
+  "image_split_resolution": null,
+  "initializer_range": 0.02,
+  "intermediate_size": 8960,
+  "llm_compress_layer_list": [
+    8,
+    16,
+    24
+  ],
+  "llm_compress_type": "attention",
+  "llm_image_token_ratio_list": [
+    1.0,
+    0.5,
+    0.25,
+    0.125
+  ],
+  "max_num_pixels": 14745600000,
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "min_slow_num_frames": 4,
+  "mm_close_init": false,
+  "mm_hidden_size": 1024,
+  "mm_llm_compress": false,
+  "mm_local_num_frames": 4,
+  "mm_newline_position": "one_token",
+  "mm_num_compress_latents": 128,
+  "mm_num_compress_query_type": "learnable",
+  "mm_patch_merge_type": "flat",
+  "mm_pos_num_frames": 4,
+  "mm_projector_lr": null,
+  "mm_projector_type": "tome16_mlp_hd64",
+  "mm_resampler_type": null,
+  "mm_spatial_pool_mode": "bilinear",
+  "mm_tunable_parts": "mm_mlp_adapter",
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -2,
+  "mm_vision_tower": "umt-hd-large",
+  "mm_vision_tower_lr": null,
+  "model_type": "qwen2",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 2,
+  "pos_skipping_range": 4096,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": true,
+  "tokenizer_model_max_length": 8192,
+  "tokenizer_padding_side": "right",
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.39.2",
+  "use_cache": true,
+  "use_mm_proj": true,
+  "use_pos_skipping": false,
+  "use_sliding_window": false,
+  "vision_encode_type": "video_image",
+  "vision_tower_pretrained": null,
+  "vocab_size": 151936
+}

stage1-UMT-Qwen2_5_2B-tome16_mlp/mm_projector.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:af5e6acfe50f50db621507be4eb184916e7107e86deffd04dc2541349af5e4b2
+size 7872568

stage1-UMT-Qwen2_5_7B_1m-tome16_mlp/config.json

@@ -0,0 +1,81 @@
+{
+  "_name_or_path": "Qwen/Qwen2.5-7B-Instruct-1M",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "dual_chunk_attention_config": {
+    "chunk_size": 262144,
+    "local_size": 8192,
+    "original_max_position_embeddings": 262144
+  },
+  "eos_token_id": 151645,
+  "frame_aspect_ratio": "square",
+  "frame_grid_pinpoints": null,
+  "hidden_act": "silu",
+  "hidden_size": 3584,
+  "image_aspect_ratio": "square",
+  "image_crop_resolution": null,
+  "image_grid_pinpoints": null,
+  "image_split_resolution": null,
+  "initializer_range": 0.02,
+  "intermediate_size": 18944,
+  "llm_compress_layer_list": [
+    8,
+    16,
+    24
+  ],
+  "llm_compress_type": "attention",
+  "llm_image_token_ratio_list": [
+    1.0,
+    0.5,
+    0.25,
+    0.125
+  ],
+  "max_num_pixels": 14745600000,
+  "max_position_embeddings": 1010000,
+  "max_window_layers": 28,
+  "min_slow_num_frames": 4,
+  "mm_close_init": false,
+  "mm_hidden_size": 1024,
+  "mm_llm_compress": false,
+  "mm_local_num_frames": 4,
+  "mm_newline_position": "one_token",
+  "mm_num_compress_latents": 128,
+  "mm_num_compress_query_type": "learnable",
+  "mm_patch_merge_type": "flat",
+  "mm_pos_num_frames": 4,
+  "mm_projector_lr": null,
+  "mm_projector_type": "tome16_mlp_hd64",
+  "mm_resampler_type": null,
+  "mm_spatial_pool_mode": "bilinear",
+  "mm_tunable_parts": "mm_mlp_adapter",
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -2,
+  "mm_vision_tower": "umt-large",
+  "mm_vision_tower_lr": null,
+  "model_type": "qwen2",
+  "num_attention_heads": 28,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 4,
+  "pos_skipping_range": 4096,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": null,
+  "rope_theta": 10000000.0,
+  "sliding_window": 32768,
+  "tie_word_embeddings": false,
+  "tokenizer_model_max_length": 8192,
+  "tokenizer_padding_side": "right",
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.39.2",
+  "use_cache": true,
+  "use_mm_proj": true,
+  "use_pos_skipping": false,
+  "use_sliding_window": false,
+  "vision_encode_type": "video_image",
+  "vision_tower_pretrained": null,
+  "vocab_size": 152064
+}

stage1-UMT-Qwen2_5_7B_1m-tome16_mlp/mm_projector.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7ec9fc9a0e530a86e891f2fbb0587ed7f689e42761bd6db3dc43f4a93a4479a
+size 33046584

stage1-UMT-Qwen2_7B-tome16_mlp/config.json

@@ -0,0 +1,75 @@
+{
+  "_name_or_path": "/mnt/petrelfs/share/videointern/Qwen2-7B-Instruct",
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 151643,
+  "eos_token_id": 151645,
+  "frame_aspect_ratio": "square",
+  "frame_grid_pinpoints": null,
+  "hidden_act": "silu",
+  "hidden_size": 3584,
+  "image_aspect_ratio": "square",
+  "image_crop_resolution": null,
+  "image_grid_pinpoints": null,
+  "image_split_resolution": null,
+  "initializer_range": 0.02,
+  "intermediate_size": 18944,
+  "llm_compress_layer_list": [
+    8,
+    16,
+    24
+  ],
+  "llm_compress_type": "attention",
+  "llm_image_token_ratio_list": [
+    1.0,
+    0.5,
+    0.25,
+    0.125
+  ],
+  "max_position_embeddings": 32768,
+  "max_window_layers": 28,
+  "min_slow_num_frames": 4,
+  "mm_close_init": false,
+  "mm_hidden_size": 1024,
+  "mm_llm_compress": false,
+  "mm_local_num_frames": 4,
+  "mm_newline_position": "one_token",
+  "mm_num_compress_latents": 128,
+  "mm_num_compress_query_type": "learnable",
+  "mm_patch_merge_type": "flat",
+  "mm_pos_num_frames": 4,
+  "mm_projector_lr": null,
+  "mm_projector_type": "tome16_mlp_hd64",
+  "mm_resampler_type": null,
+  "mm_spatial_pool_mode": "bilinear",
+  "mm_tunable_parts": "mm_mlp_adapter",
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -2,
+  "mm_vision_tower": "umt-large",
+  "mm_vision_tower_lr": null,
+  "model_type": "qwen2",
+  "num_attention_heads": 28,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 4,
+  "pos_skipping_range": 4096,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": 131072,
+  "tie_word_embeddings": false,
+  "tokenizer_model_max_length": 8192,
+  "tokenizer_padding_side": "right",
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.39.2",
+  "use_cache": true,
+  "use_mm_proj": true,
+  "use_pos_skipping": false,
+  "use_sliding_window": false,
+  "vision_encode_type": "video",
+  "vision_tower_pretrained": null,
+  "vocab_size": 152064
+}

stage1-UMT-Qwen2_7B-tome16_mlp/mm_projector.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:43c7c3ec89cb1e7090c587ac813972117ef44a6d2b94371b81212ccec89a2bed
+size 33046584