lyunee commited on 26 days ago

Commit

aaef667

verified ·

1 Parent(s): 8b45564

Upload continuous evaluation checkpoints, batch 12

Browse files

Files changed (49) hide show

.gitattributes +3 -0
continuous/qwen25/pixio_vitb16/finetune/config.json +87 -0
continuous/qwen25/pixio_vitb16/finetune/generation_config.json +13 -0
continuous/qwen25/pixio_vitb16/finetune/merges.txt +0 -0
continuous/qwen25/pixio_vitb16/finetune/model.safetensors +3 -0
continuous/qwen25/pixio_vitb16/finetune/special_tokens_map.json +31 -0
continuous/qwen25/pixio_vitb16/finetune/tokenizer.json +3 -0
continuous/qwen25/pixio_vitb16/finetune/tokenizer_config.json +208 -0
continuous/qwen25/pixio_vitb16/finetune/trainer_state.json +1856 -0
continuous/qwen25/pixio_vitb16/finetune/training_args.bin +3 -0
continuous/qwen25/pixio_vitb16/finetune/vocab.json +0 -0
continuous/qwen25/pixio_vitb16/pretrain/config.json +87 -0
continuous/qwen25/pixio_vitb16/pretrain/mm_projector.bin +3 -0
continuous/qwen25/pixio_vitb16/pretrain/trainer_state.json +421 -0
continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/added_tokens.json +24 -0
continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/chat_template.jinja +54 -0
continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/config.json +87 -0
continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/generation_config.json +13 -0
continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/merges.txt +0 -0
continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/model.safetensors +3 -0
continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/special_tokens_map.json +31 -0
continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/tokenizer.json +3 -0
continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/tokenizer_config.json +208 -0
continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/trainer_state.json +1856 -0
continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/training_args.bin +3 -0
continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/vocab.json +0 -0
continuous/qwen25/pixio_vith16/finetune/added_tokens.json +24 -0
continuous/qwen25/pixio_vith16/finetune/chat_template.jinja +54 -0
continuous/qwen25/pixio_vith16/finetune/config.json +87 -0
continuous/qwen25/pixio_vith16/finetune/generation_config.json +13 -0
continuous/qwen25/pixio_vith16/finetune/merges.txt +0 -0
continuous/qwen25/pixio_vith16/finetune/model.safetensors +3 -0
continuous/qwen25/pixio_vith16/finetune/special_tokens_map.json +31 -0
continuous/qwen25/pixio_vith16/finetune/tokenizer.json +3 -0
continuous/qwen25/pixio_vith16/finetune/tokenizer_config.json +208 -0
continuous/qwen25/pixio_vith16/finetune/trainer_state.json +1856 -0
continuous/qwen25/pixio_vith16/finetune/training_args.bin +3 -0
continuous/qwen25/pixio_vith16/finetune/vocab.json +0 -0
continuous/qwen25/pixio_vith16/pretrain.bad_hf_native_20260818/config.json +87 -0
continuous/qwen25/pixio_vith16/pretrain.bad_hf_native_20260818/mm_projector.bin +3 -0
continuous/qwen25/pixio_vith16/pretrain.bad_hf_native_20260818/trainer_state.json +421 -0
continuous/qwen25/pixio_vith16/pretrain/config.json +87 -0
continuous/qwen25/pixio_vith16/pretrain/mm_projector.bin +3 -0
continuous/qwen25/pixio_vith16/pretrain/trainer_state.json +421 -0
continuous/qwen25/pixio_vitl16/finetune/added_tokens.json +24 -0
continuous/qwen25/pixio_vitl16/finetune/chat_template.jinja +54 -0
continuous/qwen25/pixio_vitl16/finetune/config.json +87 -0
continuous/qwen25/pixio_vitl16/finetune/generation_config.json +13 -0
continuous/qwen25/pixio_vitl16/finetune/merges.txt +0 -0

.gitattributes CHANGED Viewed

@@ -84,3 +84,6 @@ continuous/qwen25/pe_lang_g14_448/finetune/tokenizer.json filter=lfs diff=lfs me
 continuous/qwen25/pe_lang_g14_448_tiling/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 continuous/qwen25/pe_lang_l14_448/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 continuous/qwen25/pixio_vit1b16/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text

 continuous/qwen25/pe_lang_g14_448_tiling/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 continuous/qwen25/pe_lang_l14_448/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 continuous/qwen25/pixio_vit1b16/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+continuous/qwen25/pixio_vitb16/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+continuous/qwen25/pixio_vith16/finetune/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/tokenizer.json filter=lfs diff=lfs merge=lfs -text

continuous/qwen25/pixio_vitb16/finetune/config.json ADDED Viewed

	@@ -0,0 +1,87 @@

+{
+  "add_faster_video": false,
+  "add_time_instruction": false,
+  "architectures": [
+    "LlavaQwenForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "dtype": "bfloat16",
+  "eos_token_id": 151645,
+  "faster_token_stride": 10,
+  "force_sample": false,
+  "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,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "mm_hidden_size": 768,
+  "mm_newline_position": "grid",
+  "mm_patch_merge_type": "flat",
+  "mm_projector_lr": 0.0001,
+  "mm_projector_type": "mlp2x_gelu",
+  "mm_resampler_type": null,
+  "mm_spatial_pool_mode": "bilinear",
+  "mm_spatial_pool_stride": null,
+  "mm_tunable_parts": "mm_mlp_adapter,mm_language_model",
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -1,
+  "mm_vision_tower": "vtb_ssl:pixio",
+  "mm_vision_tower_lr": null,
+  "model_type": "qwen2",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 2,
+  "pad_token_id": 151643,
+  "pos_skipping_range": 4096,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "tokenizer_model_max_length": 8192,
+  "tokenizer_padding_side": "right",
+  "transformers_version": "4.57.6",
+  "use_cache": true,
+  "use_mm_proj": true,
+  "use_pos_skipping": false,
+  "use_sliding_window": false,
+  "vision_tower_pretrained": "/cache/ckpt/download/tokenizer/continuous/pixio/pixio-vitb16/pixio_vitb16.pth",
+  "vocab_size": 151936
+}

continuous/qwen25/pixio_vitb16/finetune/generation_config.json ADDED Viewed

	@@ -0,0 +1,13 @@

+{
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.1,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "4.57.6"
+}

continuous/qwen25/pixio_vitb16/finetune/merges.txt ADDED Viewed

The diff for this file is too large to render. See raw diff

continuous/qwen25/pixio_vitb16/finetune/model.safetensors ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:7d1e337caaaa3338a2ae84a30e7266a06cbac4074a6e0fa2da3fcd8cf3e90e07
+size 3740113440

continuous/qwen25/pixio_vitb16/finetune/special_tokens_map.json ADDED Viewed

	@@ -0,0 +1,31 @@

+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

continuous/qwen25/pixio_vitb16/finetune/tokenizer.json ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:9c5ae00e602b8860cbd784ba82a8aa14e8feecec692e7076590d014d7b7fdafa
+size 11421896

continuous/qwen25/pixio_vitb16/finetune/tokenizer_config.json ADDED Viewed

	@@ -0,0 +1,208 @@

+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 8192,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

continuous/qwen25/pixio_vitb16/finetune/trainer_state.json ADDED Viewed

	@@ -0,0 +1,1856 @@

+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9999037443449803,
+  "eval_steps": 500,
+  "global_step": 2597,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.003850226200789296,
+      "grad_norm": 17.981914520263672,
+      "learning_rate": 1.153846153846154e-06,
+      "loss": 1.7263,
+      "step": 10
+    },
+    {
+      "epoch": 0.007700452401578592,
+      "grad_norm": 6.36418342590332,
+      "learning_rate": 2.435897435897436e-06,
+      "loss": 1.3794,
+      "step": 20
+    },
+    {
+      "epoch": 0.01155067860236789,
+      "grad_norm": 3.461688756942749,
+      "learning_rate": 3.7179487179487184e-06,
+      "loss": 1.0773,
+      "step": 30
+    },
+    {
+      "epoch": 0.015400904803157185,
+      "grad_norm": 0.7733753323554993,
+      "learning_rate": 5e-06,
+      "loss": 0.95,
+      "step": 40
+    },
+    {
+      "epoch": 0.019251131003946483,
+      "grad_norm": 0.3851611018180847,
+      "learning_rate": 6.282051282051282e-06,
+      "loss": 0.9245,
+      "step": 50
+    },
+    {
+      "epoch": 0.02310135720473578,
+      "grad_norm": 0.35280895233154297,
+      "learning_rate": 7.564102564102564e-06,
+      "loss": 0.8815,
+      "step": 60
+    },
+    {
+      "epoch": 0.026951583405525073,
+      "grad_norm": 0.2576727867126465,
+      "learning_rate": 8.846153846153847e-06,
+      "loss": 0.8575,
+      "step": 70
+    },
+    {
+      "epoch": 0.03080180960631437,
+      "grad_norm": 0.2766801118850708,
+      "learning_rate": 9.999996111488723e-06,
+      "loss": 0.8674,
+      "step": 80
+    },
+    {
+      "epoch": 0.034652035807103666,
+      "grad_norm": 0.26418715715408325,
+      "learning_rate": 9.999529497453782e-06,
+      "loss": 0.8482,
+      "step": 90
+    },
+    {
+      "epoch": 0.038502262007892966,
+      "grad_norm": 0.2665075361728668,
+      "learning_rate": 9.998285264324193e-06,
+      "loss": 0.8298,
+      "step": 100
+    },
+    {
+      "epoch": 0.04235248820868226,
+      "grad_norm": 0.23341341316699982,
+      "learning_rate": 9.996263605626052e-06,
+      "loss": 0.8181,
+      "step": 110
+    },
+    {
+      "epoch": 0.04620271440947156,
+      "grad_norm": 0.25084495544433594,
+      "learning_rate": 9.99346483580503e-06,
+      "loss": 0.8305,
+      "step": 120
+    },
+    {
+      "epoch": 0.05005294061026085,
+      "grad_norm": 0.23227085173130035,
+      "learning_rate": 9.989889390177467e-06,
+      "loss": 0.8228,
+      "step": 130
+    },
+    {
+      "epoch": 0.053903166811050146,
+      "grad_norm": 0.2537403702735901,
+      "learning_rate": 9.985537824862647e-06,
+      "loss": 0.8228,
+      "step": 140
+    },
+    {
+      "epoch": 0.057753393011839446,
+      "grad_norm": 0.25492629408836365,
+      "learning_rate": 9.980410816696317e-06,
+      "loss": 0.8061,
+      "step": 150
+    },
+    {
+      "epoch": 0.06160361921262874,
+      "grad_norm": 0.2401973158121109,
+      "learning_rate": 9.97450916312541e-06,
+      "loss": 0.8132,
+      "step": 160
+    },
+    {
+      "epoch": 0.06545384541341803,
+      "grad_norm": 0.20953841507434845,
+      "learning_rate": 9.967833782084004e-06,
+      "loss": 0.8132,
+      "step": 170
+    },
+    {
+      "epoch": 0.06930407161420733,
+      "grad_norm": 0.2545095384120941,
+      "learning_rate": 9.960385711850554e-06,
+      "loss": 0.8417,
+      "step": 180
+    },
+    {
+      "epoch": 0.07315429781499663,
+      "grad_norm": 0.2435549795627594,
+      "learning_rate": 9.952166110886394e-06,
+      "loss": 0.8316,
+      "step": 190
+    },
+    {
+      "epoch": 0.07700452401578593,
+      "grad_norm": 0.2697771191596985,
+      "learning_rate": 9.943176257655567e-06,
+      "loss": 0.8255,
+      "step": 200
+    },
+    {
+      "epoch": 0.08085475021657522,
+      "grad_norm": 0.23707957565784454,
+      "learning_rate": 9.933417550425952e-06,
+      "loss": 0.8142,
+      "step": 210
+    },
+    {
+      "epoch": 0.08470497641736452,
+      "grad_norm": 0.26199057698249817,
+      "learning_rate": 9.9228915070518e-06,
+      "loss": 0.8042,
+      "step": 220
+    },
+    {
+      "epoch": 0.08855520261815382,
+      "grad_norm": 0.24625518918037415,
+      "learning_rate": 9.911599764737632e-06,
+      "loss": 0.8065,
+      "step": 230
+    },
+    {
+      "epoch": 0.09240542881894312,
+      "grad_norm": 0.2566681206226349,
+      "learning_rate": 9.899544079783606e-06,
+      "loss": 0.8004,
+      "step": 240
+    },
+    {
+      "epoch": 0.0962556550197324,
+      "grad_norm": 0.24434499442577362,
+      "learning_rate": 9.886726327312338e-06,
+      "loss": 0.8154,
+      "step": 250
+    },
+    {
+      "epoch": 0.1001058812205217,
+      "grad_norm": 0.2511313259601593,
+      "learning_rate": 9.87314850097724e-06,
+      "loss": 0.8151,
+      "step": 260
+    },
+    {
+      "epoch": 0.103956107421311,
+      "grad_norm": 0.2406085729598999,
+      "learning_rate": 9.858812712652452e-06,
+      "loss": 0.8035,
+      "step": 270
+    },
+    {
+      "epoch": 0.10780633362210029,
+      "grad_norm": 0.2263503074645996,
+      "learning_rate": 9.843721192104341e-06,
+      "loss": 0.8038,
+      "step": 280
+    },
+    {
+      "epoch": 0.11165655982288959,
+      "grad_norm": 0.23683714866638184,
+      "learning_rate": 9.827876286644698e-06,
+      "loss": 0.8236,
+      "step": 290
+    },
+    {
+      "epoch": 0.11550678602367889,
+      "grad_norm": 0.21940334141254425,
+      "learning_rate": 9.811280460765643e-06,
+      "loss": 0.7841,
+      "step": 300
+    },
+    {
+      "epoch": 0.11935701222446819,
+      "grad_norm": 0.232316792011261,
+      "learning_rate": 9.793936295756292e-06,
+      "loss": 0.7939,
+      "step": 310
+    },
+    {
+      "epoch": 0.12320723842525748,
+      "grad_norm": 0.2401176542043686,
+      "learning_rate": 9.775846489301279e-06,
+      "loss": 0.7983,
+      "step": 320
+    },
+    {
+      "epoch": 0.12705746462604678,
+      "grad_norm": 0.23774683475494385,
+      "learning_rate": 9.757013855061157e-06,
+      "loss": 0.7961,
+      "step": 330
+    },
+    {
+      "epoch": 0.13090769082683607,
+      "grad_norm": 0.24004606902599335,
+      "learning_rate": 9.737441322234758e-06,
+      "loss": 0.8098,
+      "step": 340
+    },
+    {
+      "epoch": 0.13475791702762538,
+      "grad_norm": 0.2503903806209564,
+      "learning_rate": 9.717131935103605e-06,
+      "loss": 0.808,
+      "step": 350
+    },
+    {
+      "epoch": 0.13860814322841467,
+      "grad_norm": 0.23993006348609924,
+      "learning_rate": 9.696088852558398e-06,
+      "loss": 0.7938,
+      "step": 360
+    },
+    {
+      "epoch": 0.14245836942920398,
+      "grad_norm": 0.23292550444602966,
+      "learning_rate": 9.674315347607686e-06,
+      "loss": 0.8196,
+      "step": 370
+    },
+    {
+      "epoch": 0.14630859562999327,
+      "grad_norm": 0.2299531251192093,
+      "learning_rate": 9.651814806868796e-06,
+      "loss": 0.7848,
+      "step": 380
+    },
+    {
+      "epoch": 0.15015882183078255,
+      "grad_norm": 0.2313643842935562,
+      "learning_rate": 9.62859073004107e-06,
+      "loss": 0.7896,
+      "step": 390
+    },
+    {
+      "epoch": 0.15400904803157187,
+      "grad_norm": 0.21348530054092407,
+      "learning_rate": 9.604646729361547e-06,
+      "loss": 0.7962,
+      "step": 400
+    },
+    {
+      "epoch": 0.15785927423236115,
+      "grad_norm": 0.22961300611495972,
+      "learning_rate": 9.579986529043099e-06,
+      "loss": 0.7881,
+      "step": 410
+    },
+    {
+      "epoch": 0.16170950043315044,
+      "grad_norm": 0.24253366887569427,
+      "learning_rate": 9.554613964695189e-06,
+      "loss": 0.8001,
+      "step": 420
+    },
+    {
+      "epoch": 0.16555972663393975,
+      "grad_norm": 0.23166470229625702,
+      "learning_rate": 9.528532982727281e-06,
+      "loss": 0.791,
+      "step": 430
+    },
+    {
+      "epoch": 0.16940995283472904,
+      "grad_norm": 0.22402088344097137,
+      "learning_rate": 9.501747639735018e-06,
+      "loss": 0.7903,
+      "step": 440
+    },
+    {
+      "epoch": 0.17326017903551832,
+      "grad_norm": 0.23103733360767365,
+      "learning_rate": 9.474262101869272e-06,
+      "loss": 0.7787,
+      "step": 450
+    },
+    {
+      "epoch": 0.17711040523630764,
+      "grad_norm": 0.22975780069828033,
+      "learning_rate": 9.446080644188141e-06,
+      "loss": 0.7836,
+      "step": 460
+    },
+    {
+      "epoch": 0.18096063143709692,
+      "grad_norm": 0.21977214515209198,
+      "learning_rate": 9.417207649992017e-06,
+      "loss": 0.7925,
+      "step": 470
+    },
+    {
+      "epoch": 0.18481085763788624,
+      "grad_norm": 0.23165468871593475,
+      "learning_rate": 9.387647610141816e-06,
+      "loss": 0.7921,
+      "step": 480
+    },
+    {
+      "epoch": 0.18866108383867553,
+      "grad_norm": 0.21952630579471588,
+      "learning_rate": 9.35740512236047e-06,
+      "loss": 0.7771,
+      "step": 490
+    },
+    {
+      "epoch": 0.1925113100394648,
+      "grad_norm": 0.20784978568553925,
+      "learning_rate": 9.326484890517809e-06,
+      "loss": 0.7855,
+      "step": 500
+    },
+    {
+      "epoch": 0.19636153624025413,
+      "grad_norm": 0.23287734389305115,
+      "learning_rate": 9.294891723898926e-06,
+      "loss": 0.7772,
+      "step": 510
+    },
+    {
+      "epoch": 0.2002117624410434,
+      "grad_norm": 0.24799081683158875,
+      "learning_rate": 9.262630536456153e-06,
+      "loss": 0.8066,
+      "step": 520
+    },
+    {
+      "epoch": 0.2040619886418327,
+      "grad_norm": 0.24101607501506805,
+      "learning_rate": 9.229706346044749e-06,
+      "loss": 0.7824,
+      "step": 530
+    },
+    {
+      "epoch": 0.207912214842622,
+      "grad_norm": 0.2474917471408844,
+      "learning_rate": 9.196124273642422e-06,
+      "loss": 0.7923,
+      "step": 540
+    },
+    {
+      "epoch": 0.2117624410434113,
+      "grad_norm": 0.234390527009964,
+      "learning_rate": 9.161889542552842e-06,
+      "loss": 0.7947,
+      "step": 550
+    },
+    {
+      "epoch": 0.21561266724420058,
+      "grad_norm": 0.22415609657764435,
+      "learning_rate": 9.127007477593193e-06,
+      "loss": 0.8003,
+      "step": 560
+    },
+    {
+      "epoch": 0.2194628934449899,
+      "grad_norm": 0.24675478041172028,
+      "learning_rate": 9.091483504265974e-06,
+      "loss": 0.8106,
+      "step": 570
+    },
+    {
+      "epoch": 0.22331311964577918,
+      "grad_norm": 0.23582714796066284,
+      "learning_rate": 9.055323147915116e-06,
+      "loss": 0.7875,
+      "step": 580
+    },
+    {
+      "epoch": 0.2271633458465685,
+      "grad_norm": 0.23336561024188995,
+      "learning_rate": 9.018532032866585e-06,
+      "loss": 0.7995,
+      "step": 590
+    },
+    {
+      "epoch": 0.23101357204735778,
+      "grad_norm": 0.2259475141763687,
+      "learning_rate": 8.981115881553579e-06,
+      "loss": 0.7909,
+      "step": 600
+    },
+    {
+      "epoch": 0.23486379824814707,
+      "grad_norm": 0.23881524801254272,
+      "learning_rate": 8.943080513626476e-06,
+      "loss": 0.8109,
+      "step": 610
+    },
+    {
+      "epoch": 0.23871402444893638,
+      "grad_norm": 0.22244426608085632,
+      "learning_rate": 8.904431845047641e-06,
+      "loss": 0.7953,
+      "step": 620
+    },
+    {
+      "epoch": 0.24256425064972567,
+      "grad_norm": 0.24727030098438263,
+      "learning_rate": 8.86517588717128e-06,
+      "loss": 0.8113,
+      "step": 630
+    },
+    {
+      "epoch": 0.24641447685051496,
+      "grad_norm": 0.21751193702220917,
+      "learning_rate": 8.82531874580844e-06,
+      "loss": 0.8004,
+      "step": 640
+    },
+    {
+      "epoch": 0.25026470305130427,
+      "grad_norm": 0.2105468511581421,
+      "learning_rate": 8.784866620277314e-06,
+      "loss": 0.7581,
+      "step": 650
+    },
+    {
+      "epoch": 0.25411492925209356,
+      "grad_norm": 0.2531110942363739,
+      "learning_rate": 8.743825802439018e-06,
+      "loss": 0.7951,
+      "step": 660
+    },
+    {
+      "epoch": 0.25796515545288284,
+      "grad_norm": 0.24918149411678314,
+      "learning_rate": 8.702202675718956e-06,
+      "loss": 0.7992,
+      "step": 670
+    },
+    {
+      "epoch": 0.26181538165367213,
+      "grad_norm": 0.2251148670911789,
+      "learning_rate": 8.660003714113959e-06,
+      "loss": 0.7696,
+      "step": 680
+    },
+    {
+      "epoch": 0.26566560785446147,
+      "grad_norm": 0.22266755998134613,
+      "learning_rate": 8.617235481185323e-06,
+      "loss": 0.7936,
+      "step": 690
+    },
+    {
+      "epoch": 0.26951583405525076,
+      "grad_norm": 0.23317226767539978,
+      "learning_rate": 8.573904629037931e-06,
+      "loss": 0.7971,
+      "step": 700
+    },
+    {
+      "epoch": 0.27336606025604004,
+      "grad_norm": 0.22407065331935883,
+      "learning_rate": 8.530017897285585e-06,
+      "loss": 0.7786,
+      "step": 710
+    },
+    {
+      "epoch": 0.27721628645682933,
+      "grad_norm": 0.21442089974880219,
+      "learning_rate": 8.485582112002746e-06,
+      "loss": 0.7836,
+      "step": 720
+    },
+    {
+      "epoch": 0.2810665126576186,
+      "grad_norm": 0.2258174866437912,
+      "learning_rate": 8.440604184662811e-06,
+      "loss": 0.7882,
+      "step": 730
+    },
+    {
+      "epoch": 0.28491673885840796,
+      "grad_norm": 0.2330341339111328,
+      "learning_rate": 8.395091111063118e-06,
+      "loss": 0.7787,
+      "step": 740
+    },
+    {
+      "epoch": 0.28876696505919724,
+      "grad_norm": 0.21531911194324493,
+      "learning_rate": 8.349049970236822e-06,
+      "loss": 0.7768,
+      "step": 750
+    },
+    {
+      "epoch": 0.29261719125998653,
+      "grad_norm": 0.22701889276504517,
+      "learning_rate": 8.302487923351842e-06,
+      "loss": 0.7698,
+      "step": 760
+    },
+    {
+      "epoch": 0.2964674174607758,
+      "grad_norm": 0.217278391122818,
+      "learning_rate": 8.255412212597021e-06,
+      "loss": 0.7867,
+      "step": 770
+    },
+    {
+      "epoch": 0.3003176436615651,
+      "grad_norm": 0.2132900208234787,
+      "learning_rate": 8.207830160055688e-06,
+      "loss": 0.7826,
+      "step": 780
+    },
+    {
+      "epoch": 0.3041678698623544,
+      "grad_norm": 0.23059262335300446,
+      "learning_rate": 8.159749166566786e-06,
+      "loss": 0.7703,
+      "step": 790
+    },
+    {
+      "epoch": 0.30801809606314373,
+      "grad_norm": 0.2308778017759323,
+      "learning_rate": 8.11117671057377e-06,
+      "loss": 0.7968,
+      "step": 800
+    },
+    {
+      "epoch": 0.311868322263933,
+      "grad_norm": 0.24685069918632507,
+      "learning_rate": 8.062120346961404e-06,
+      "loss": 0.7893,
+      "step": 810
+    },
+    {
+      "epoch": 0.3157185484647223,
+      "grad_norm": 0.21524527668952942,
+      "learning_rate": 8.012587705880705e-06,
+      "loss": 0.7841,
+      "step": 820
+    },
+    {
+      "epoch": 0.3195687746655116,
+      "grad_norm": 0.2267124503850937,
+      "learning_rate": 7.962586491562146e-06,
+      "loss": 0.7699,
+      "step": 830
+    },
+    {
+      "epoch": 0.3234190008663009,
+      "grad_norm": 0.24558836221694946,
+      "learning_rate": 7.912124481117357e-06,
+      "loss": 0.7748,
+      "step": 840
+    },
+    {
+      "epoch": 0.3272692270670902,
+      "grad_norm": 0.2220401018857956,
+      "learning_rate": 7.861209523329494e-06,
+      "loss": 0.7846,
+      "step": 850
+    },
+    {
+      "epoch": 0.3311194532678795,
+      "grad_norm": 0.22430536150932312,
+      "learning_rate": 7.809849537432432e-06,
+      "loss": 0.7714,
+      "step": 860
+    },
+    {
+      "epoch": 0.3349696794686688,
+      "grad_norm": 0.2187328040599823,
+      "learning_rate": 7.758052511879035e-06,
+      "loss": 0.7797,
+      "step": 870
+    },
+    {
+      "epoch": 0.3388199056694581,
+      "grad_norm": 0.2099825143814087,
+      "learning_rate": 7.705826503098642e-06,
+      "loss": 0.7745,
+      "step": 880
+    },
+    {
+      "epoch": 0.34267013187024736,
+      "grad_norm": 0.21443051099777222,
+      "learning_rate": 7.653179634243979e-06,
+      "loss": 0.7742,
+      "step": 890
+    },
+    {
+      "epoch": 0.34652035807103665,
+      "grad_norm": 0.20142491161823273,
+      "learning_rate": 7.600120093927696e-06,
+      "loss": 0.78,
+      "step": 900
+    },
+    {
+      "epoch": 0.350370584271826,
+      "grad_norm": 0.25082746148109436,
+      "learning_rate": 7.546656134948729e-06,
+      "loss": 0.7654,
+      "step": 910
+    },
+    {
+      "epoch": 0.3542208104726153,
+      "grad_norm": 0.24113531410694122,
+      "learning_rate": 7.492796073008661e-06,
+      "loss": 0.7666,
+      "step": 920
+    },
+    {
+      "epoch": 0.35807103667340456,
+      "grad_norm": 0.24717530608177185,
+      "learning_rate": 7.438548285418326e-06,
+      "loss": 0.7555,
+      "step": 930
+    },
+    {
+      "epoch": 0.36192126287419385,
+      "grad_norm": 0.2578974664211273,
+      "learning_rate": 7.3839212097948e-06,
+      "loss": 0.7874,
+      "step": 940
+    },
+    {
+      "epoch": 0.36577148907498314,
+      "grad_norm": 0.23232412338256836,
+      "learning_rate": 7.328923342749039e-06,
+      "loss": 0.8014,
+      "step": 950
+    },
+    {
+      "epoch": 0.3696217152757725,
+      "grad_norm": 0.23840868473052979,
+      "learning_rate": 7.273563238564319e-06,
+      "loss": 0.7877,
+      "step": 960
+    },
+    {
+      "epoch": 0.37347194147656176,
+      "grad_norm": 0.23213908076286316,
+      "learning_rate": 7.217849507865724e-06,
+      "loss": 0.7678,
+      "step": 970
+    },
+    {
+      "epoch": 0.37732216767735105,
+      "grad_norm": 0.2209821343421936,
+      "learning_rate": 7.161790816280858e-06,
+      "loss": 0.7725,
+      "step": 980
+    },
+    {
+      "epoch": 0.38117239387814034,
+      "grad_norm": 0.23177284002304077,
+      "learning_rate": 7.1053958830920125e-06,
+      "loss": 0.7589,
+      "step": 990
+    },
+    {
+      "epoch": 0.3850226200789296,
+      "grad_norm": 0.2209496647119522,
+      "learning_rate": 7.0486734798799715e-06,
+      "loss": 0.7716,
+      "step": 1000
+    },
+    {
+      "epoch": 0.3888728462797189,
+      "grad_norm": 0.2186221480369568,
+      "learning_rate": 6.991632429159708e-06,
+      "loss": 0.7973,
+      "step": 1010
+    },
+    {
+      "epoch": 0.39272307248050825,
+      "grad_norm": 0.24343648552894592,
+      "learning_rate": 6.934281603008132e-06,
+      "loss": 0.7907,
+      "step": 1020
+    },
+    {
+      "epoch": 0.39657329868129754,
+      "grad_norm": 0.22137701511383057,
+      "learning_rate": 6.876629921684143e-06,
+      "loss": 0.7875,
+      "step": 1030
+    },
+    {
+      "epoch": 0.4004235248820868,
+      "grad_norm": 0.2306416779756546,
+      "learning_rate": 6.818686352241199e-06,
+      "loss": 0.7783,
+      "step": 1040
+    },
+    {
+      "epoch": 0.4042737510828761,
+      "grad_norm": 0.21005044877529144,
+      "learning_rate": 6.760459907132573e-06,
+      "loss": 0.7746,
+      "step": 1050
+    },
+    {
+      "epoch": 0.4081239772836654,
+      "grad_norm": 0.21057558059692383,
+      "learning_rate": 6.701959642809589e-06,
+      "loss": 0.7812,
+      "step": 1060
+    },
+    {
+      "epoch": 0.41197420348445474,
+      "grad_norm": 0.22203101217746735,
+      "learning_rate": 6.643194658312987e-06,
+      "loss": 0.7709,
+      "step": 1070
+    },
+    {
+      "epoch": 0.415824429685244,
+      "grad_norm": 0.22568464279174805,
+      "learning_rate": 6.584174093857676e-06,
+      "loss": 0.7954,
+      "step": 1080
+    },
+    {
+      "epoch": 0.4196746558860333,
+      "grad_norm": 0.25083810091018677,
+      "learning_rate": 6.5249071294110735e-06,
+      "loss": 0.7779,
+      "step": 1090
+    },
+    {
+      "epoch": 0.4235248820868226,
+      "grad_norm": 0.23170016705989838,
+      "learning_rate": 6.465402983265274e-06,
+      "loss": 0.792,
+      "step": 1100
+    },
+    {
+      "epoch": 0.4273751082876119,
+      "grad_norm": 0.23728950321674347,
+      "learning_rate": 6.405670910603243e-06,
+      "loss": 0.788,
+      "step": 1110
+    },
+    {
+      "epoch": 0.43122533448840117,
+      "grad_norm": 0.25516369938850403,
+      "learning_rate": 6.345720202059291e-06,
+      "loss": 0.7841,
+      "step": 1120
+    },
+    {
+      "epoch": 0.4350755606891905,
+      "grad_norm": 0.22734245657920837,
+      "learning_rate": 6.285560182274008e-06,
+      "loss": 0.7622,
+      "step": 1130
+    },
+    {
+      "epoch": 0.4389257868899798,
+      "grad_norm": 0.2242775857448578,
+      "learning_rate": 6.225200208443936e-06,
+      "loss": 0.7822,
+      "step": 1140
+    },
+    {
+      "epoch": 0.4427760130907691,
+      "grad_norm": 0.22765618562698364,
+      "learning_rate": 6.16464966886616e-06,
+      "loss": 0.7752,
+      "step": 1150
+    },
+    {
+      "epoch": 0.44662623929155837,
+      "grad_norm": 0.26240643858909607,
+      "learning_rate": 6.103917981478065e-06,
+      "loss": 0.7976,
+      "step": 1160
+    },
+    {
+      "epoch": 0.45047646549234766,
+      "grad_norm": 0.22381991147994995,
+      "learning_rate": 6.043014592392489e-06,
+      "loss": 0.8098,
+      "step": 1170
+    },
+    {
+      "epoch": 0.454326691693137,
+      "grad_norm": 0.208811417222023,
+      "learning_rate": 5.981948974428485e-06,
+      "loss": 0.7821,
+      "step": 1180
+    },
+    {
+      "epoch": 0.4581769178939263,
+      "grad_norm": 0.23448197543621063,
+      "learning_rate": 5.920730625637934e-06,
+      "loss": 0.7639,
+      "step": 1190
+    },
+    {
+      "epoch": 0.46202714409471557,
+      "grad_norm": 0.236393541097641,
+      "learning_rate": 5.8593690678282355e-06,
+      "loss": 0.7747,
+      "step": 1200
+    },
+    {
+      "epoch": 0.46587737029550486,
+      "grad_norm": 0.2450762391090393,
+      "learning_rate": 5.797873845081301e-06,
+      "loss": 0.7764,
+      "step": 1210
+    },
+    {
+      "epoch": 0.46972759649629414,
+      "grad_norm": 0.22729001939296722,
+      "learning_rate": 5.736254522269075e-06,
+      "loss": 0.7935,
+      "step": 1220
+    },
+    {
+      "epoch": 0.47357782269708343,
+      "grad_norm": 0.21898922324180603,
+      "learning_rate": 5.6745206835658386e-06,
+      "loss": 0.7971,
+      "step": 1230
+    },
+    {
+      "epoch": 0.47742804889787277,
+      "grad_norm": 0.21546868979930878,
+      "learning_rate": 5.612681930957495e-06,
+      "loss": 0.765,
+      "step": 1240
+    },
+    {
+      "epoch": 0.48127827509866206,
+      "grad_norm": 0.21499834954738617,
+      "learning_rate": 5.550747882748095e-06,
+      "loss": 0.7643,
+      "step": 1250
+    },
+    {
+      "epoch": 0.48512850129945134,
+      "grad_norm": 0.24675637483596802,
+      "learning_rate": 5.4887281720638195e-06,
+      "loss": 0.7718,
+      "step": 1260
+    },
+    {
+      "epoch": 0.48897872750024063,
+      "grad_norm": 0.2622359097003937,
+      "learning_rate": 5.42663244535466e-06,
+      "loss": 0.7579,
+      "step": 1270
+    },
+    {
+      "epoch": 0.4928289537010299,
+      "grad_norm": 0.22664067149162292,
+      "learning_rate": 5.3644703608940205e-06,
+      "loss": 0.7649,
+      "step": 1280
+    },
+    {
+      "epoch": 0.49667917990181926,
+      "grad_norm": 0.2359621822834015,
+      "learning_rate": 5.302251587276491e-06,
+      "loss": 0.7756,
+      "step": 1290
+    },
+    {
+      "epoch": 0.5005294061026085,
+      "grad_norm": 0.21736198663711548,
+      "learning_rate": 5.2399858019140005e-06,
+      "loss": 0.7773,
+      "step": 1300
+    },
+    {
+      "epoch": 0.5043796323033978,
+      "grad_norm": 0.2143339365720749,
+      "learning_rate": 5.177682689530615e-06,
+      "loss": 0.7879,
+      "step": 1310
+    },
+    {
+      "epoch": 0.5082298585041871,
+      "grad_norm": 0.23065048456192017,
+      "learning_rate": 5.11535194065619e-06,
+      "loss": 0.7608,
+      "step": 1320
+    },
+    {
+      "epoch": 0.5120800847049765,
+      "grad_norm": 0.25001251697540283,
+      "learning_rate": 5.053003250119113e-06,
+      "loss": 0.7699,
+      "step": 1330
+    },
+    {
+      "epoch": 0.5159303109057657,
+      "grad_norm": 0.23537087440490723,
+      "learning_rate": 4.9906463155383925e-06,
+      "loss": 0.7746,
+      "step": 1340
+    },
+    {
+      "epoch": 0.519780537106555,
+      "grad_norm": 0.22238434851169586,
+      "learning_rate": 4.9282908358153046e-06,
+      "loss": 0.7661,
+      "step": 1350
+    },
+    {
+      "epoch": 0.5236307633073443,
+      "grad_norm": 0.25297248363494873,
+      "learning_rate": 4.8659465096248345e-06,
+      "loss": 0.7866,
+      "step": 1360
+    },
+    {
+      "epoch": 0.5274809895081336,
+      "grad_norm": 0.233872652053833,
+      "learning_rate": 4.803623033907169e-06,
+      "loss": 0.7849,
+      "step": 1370
+    },
+    {
+      "epoch": 0.5313312157089229,
+      "grad_norm": 0.2154068648815155,
+      "learning_rate": 4.741330102359441e-06,
+      "loss": 0.7905,
+      "step": 1380
+    },
+    {
+      "epoch": 0.5351814419097122,
+      "grad_norm": 0.2389075607061386,
+      "learning_rate": 4.679077403927996e-06,
+      "loss": 0.7701,
+      "step": 1390
+    },
+    {
+      "epoch": 0.5390316681105015,
+      "grad_norm": 0.22233641147613525,
+      "learning_rate": 4.616874621301373e-06,
+      "loss": 0.7697,
+      "step": 1400
+    },
+    {
+      "epoch": 0.5428818943112907,
+      "grad_norm": 0.24262259900569916,
+      "learning_rate": 4.554731429404293e-06,
+      "loss": 0.784,
+      "step": 1410
+    },
+    {
+      "epoch": 0.5467321205120801,
+      "grad_norm": 0.23879680037498474,
+      "learning_rate": 4.492657493892822e-06,
+      "loss": 0.7703,
+      "step": 1420
+    },
+    {
+      "epoch": 0.5505823467128694,
+      "grad_norm": 0.23875769972801208,
+      "learning_rate": 4.430662469650999e-06,
+      "loss": 0.7853,
+      "step": 1430
+    },
+    {
+      "epoch": 0.5544325729136587,
+      "grad_norm": 0.2339104413986206,
+      "learning_rate": 4.368755999289118e-06,
+      "loss": 0.7884,
+      "step": 1440
+    },
+    {
+      "epoch": 0.558282799114448,
+      "grad_norm": 0.24357962608337402,
+      "learning_rate": 4.306947711643949e-06,
+      "loss": 0.7705,
+      "step": 1450
+    },
+    {
+      "epoch": 0.5621330253152372,
+      "grad_norm": 0.23774825036525726,
+      "learning_rate": 4.245247220281068e-06,
+      "loss": 0.7752,
+      "step": 1460
+    },
+    {
+      "epoch": 0.5659832515160266,
+      "grad_norm": 0.22439587116241455,
+      "learning_rate": 4.183664121999587e-06,
+      "loss": 0.7719,
+      "step": 1470
+    },
+    {
+      "epoch": 0.5698334777168159,
+      "grad_norm": 0.2343369424343109,
+      "learning_rate": 4.122207995339474e-06,
+      "loss": 0.7676,
+      "step": 1480
+    },
+    {
+      "epoch": 0.5736837039176051,
+      "grad_norm": 0.23909428715705872,
+      "learning_rate": 4.060888399091734e-06,
+      "loss": 0.7738,
+      "step": 1490
+    },
+    {
+      "epoch": 0.5775339301183945,
+      "grad_norm": 0.21251186728477478,
+      "learning_rate": 3.999714870811639e-06,
+      "loss": 0.7827,
+      "step": 1500
+    },
+    {
+      "epoch": 0.5813841563191837,
+      "grad_norm": 0.22653928399085999,
+      "learning_rate": 3.938696925335275e-06,
+      "loss": 0.7764,
+      "step": 1510
+    },
+    {
+      "epoch": 0.5852343825199731,
+      "grad_norm": 0.2305472195148468,
+      "learning_rate": 3.87784405329962e-06,
+      "loss": 0.7853,
+      "step": 1520
+    },
+    {
+      "epoch": 0.5890846087207623,
+      "grad_norm": 0.2225639671087265,
+      "learning_rate": 3.817165719666382e-06,
+      "loss": 0.7874,
+      "step": 1530
+    },
+    {
+      "epoch": 0.5929348349215516,
+      "grad_norm": 0.22117744386196136,
+      "learning_rate": 3.756671362249841e-06,
+      "loss": 0.7905,
+      "step": 1540
+    },
+    {
+      "epoch": 0.596785061122341,
+      "grad_norm": 0.22345326840877533,
+      "learning_rate": 3.6963703902488907e-06,
+      "loss": 0.7776,
+      "step": 1550
+    },
+    {
+      "epoch": 0.6006352873231302,
+      "grad_norm": 0.22852644324302673,
+      "learning_rate": 3.6362721827835655e-06,
+      "loss": 0.7949,
+      "step": 1560
+    },
+    {
+      "epoch": 0.6044855135239195,
+      "grad_norm": 0.23226700723171234,
+      "learning_rate": 3.576386087436212e-06,
+      "loss": 0.7893,
+      "step": 1570
+    },
+    {
+      "epoch": 0.6083357397247088,
+      "grad_norm": 0.2426396757364273,
+      "learning_rate": 3.5167214187975913e-06,
+      "loss": 0.7789,
+      "step": 1580
+    },
+    {
+      "epoch": 0.6121859659254981,
+      "grad_norm": 0.20419225096702576,
+      "learning_rate": 3.457287457018086e-06,
+      "loss": 0.7803,
+      "step": 1590
+    },
+    {
+      "epoch": 0.6160361921262875,
+      "grad_norm": 0.22597230970859528,
+      "learning_rate": 3.3980934463643003e-06,
+      "loss": 0.7692,
+      "step": 1600
+    },
+    {
+      "epoch": 0.6198864183270767,
+      "grad_norm": 0.21259953081607819,
+      "learning_rate": 3.3391485937812053e-06,
+      "loss": 0.7499,
+      "step": 1610
+    },
+    {
+      "epoch": 0.623736644527866,
+      "grad_norm": 0.22929610311985016,
+      "learning_rate": 3.280462067460115e-06,
+      "loss": 0.7505,
+      "step": 1620
+    },
+    {
+      "epoch": 0.6275868707286553,
+      "grad_norm": 0.22465169429779053,
+      "learning_rate": 3.222042995412669e-06,
+      "loss": 0.7821,
+      "step": 1630
+    },
+    {
+      "epoch": 0.6314370969294446,
+      "grad_norm": 0.20021402835845947,
+      "learning_rate": 3.163900464051089e-06,
+      "loss": 0.7704,
+      "step": 1640
+    },
+    {
+      "epoch": 0.635287323130234,
+      "grad_norm": 0.22863319516181946,
+      "learning_rate": 3.1060435167748826e-06,
+      "loss": 0.7801,
+      "step": 1650
+    },
+    {
+      "epoch": 0.6391375493310232,
+      "grad_norm": 0.22787386178970337,
+      "learning_rate": 3.048481152564253e-06,
+      "loss": 0.7601,
+      "step": 1660
+    },
+    {
+      "epoch": 0.6429877755318125,
+      "grad_norm": 0.2464853823184967,
+      "learning_rate": 2.9912223245804017e-06,
+      "loss": 0.7794,
+      "step": 1670
+    },
+    {
+      "epoch": 0.6468380017326018,
+      "grad_norm": 0.2368755340576172,
+      "learning_rate": 2.934275938772981e-06,
+      "loss": 0.7772,
+      "step": 1680
+    },
+    {
+      "epoch": 0.6506882279333911,
+      "grad_norm": 0.23290333151817322,
+      "learning_rate": 2.8776508524948676e-06,
+      "loss": 0.7667,
+      "step": 1690
+    },
+    {
+      "epoch": 0.6545384541341804,
+      "grad_norm": 0.2066739946603775,
+      "learning_rate": 2.821355873124503e-06,
+      "loss": 0.7681,
+      "step": 1700
+    },
+    {
+      "epoch": 0.6583886803349697,
+      "grad_norm": 0.2349947690963745,
+      "learning_rate": 2.7653997566960144e-06,
+      "loss": 0.7791,
+      "step": 1710
+    },
+    {
+      "epoch": 0.662238906535759,
+      "grad_norm": 0.23560185730457306,
+      "learning_rate": 2.709791206537309e-06,
+      "loss": 0.7688,
+      "step": 1720
+    },
+    {
+      "epoch": 0.6660891327365482,
+      "grad_norm": 0.22887460887432098,
+      "learning_rate": 2.6545388719163747e-06,
+      "loss": 0.7756,
+      "step": 1730
+    },
+    {
+      "epoch": 0.6699393589373376,
+      "grad_norm": 0.2311510145664215,
+      "learning_rate": 2.599651346695979e-06,
+      "loss": 0.7737,
+      "step": 1740
+    },
+    {
+      "epoch": 0.6737895851381268,
+      "grad_norm": 0.21698802709579468,
+      "learning_rate": 2.545137167997007e-06,
+      "loss": 0.7717,
+      "step": 1750
+    },
+    {
+      "epoch": 0.6776398113389162,
+      "grad_norm": 0.21525710821151733,
+      "learning_rate": 2.4910048148705972e-06,
+      "loss": 0.7614,
+      "step": 1760
+    },
+    {
+      "epoch": 0.6814900375397055,
+      "grad_norm": 0.23916399478912354,
+      "learning_rate": 2.4372627069793327e-06,
+      "loss": 0.8032,
+      "step": 1770
+    },
+    {
+      "epoch": 0.6853402637404947,
+      "grad_norm": 0.23879562318325043,
+      "learning_rate": 2.3839192032876486e-06,
+      "loss": 0.7898,
+      "step": 1780
+    },
+    {
+      "epoch": 0.6891904899412841,
+      "grad_norm": 0.22668789327144623,
+      "learning_rate": 2.3309826007617047e-06,
+      "loss": 0.7995,
+      "step": 1790
+    },
+    {
+      "epoch": 0.6930407161420733,
+      "grad_norm": 0.22550971806049347,
+      "learning_rate": 2.278461133078881e-06,
+      "loss": 0.7632,
+      "step": 1800
+    },
+    {
+      "epoch": 0.6968909423428626,
+      "grad_norm": 0.23883166909217834,
+      "learning_rate": 2.2263629693471258e-06,
+      "loss": 0.755,
+      "step": 1810
+    },
+    {
+      "epoch": 0.700741168543652,
+      "grad_norm": 0.2183171659708023,
+      "learning_rate": 2.1746962128343414e-06,
+      "loss": 0.7775,
+      "step": 1820
+    },
+    {
+      "epoch": 0.7045913947444412,
+      "grad_norm": 0.23084194958209991,
+      "learning_rate": 2.1234688997080183e-06,
+      "loss": 0.7673,
+      "step": 1830
+    },
+    {
+      "epoch": 0.7084416209452306,
+      "grad_norm": 0.23886601626873016,
+      "learning_rate": 2.0726889977853053e-06,
+      "loss": 0.7735,
+      "step": 1840
+    },
+    {
+      "epoch": 0.7122918471460198,
+      "grad_norm": 0.24208693206310272,
+      "learning_rate": 2.022364405293703e-06,
+      "loss": 0.7829,
+      "step": 1850
+    },
+    {
+      "epoch": 0.7161420733468091,
+      "grad_norm": 0.223402202129364,
+      "learning_rate": 1.9725029496425814e-06,
+      "loss": 0.7786,
+      "step": 1860
+    },
+    {
+      "epoch": 0.7199922995475985,
+      "grad_norm": 0.2074689418077469,
+      "learning_rate": 1.923112386205727e-06,
+      "loss": 0.7738,
+      "step": 1870
+    },
+    {
+      "epoch": 0.7238425257483877,
+      "grad_norm": 0.21802689135074615,
+      "learning_rate": 1.874200397115079e-06,
+      "loss": 0.7817,
+      "step": 1880
+    },
+    {
+      "epoch": 0.727692751949177,
+      "grad_norm": 0.23853886127471924,
+      "learning_rate": 1.825774590065857e-06,
+      "loss": 0.7644,
+      "step": 1890
+    },
+    {
+      "epoch": 0.7315429781499663,
+      "grad_norm": 0.2347031831741333,
+      "learning_rate": 1.7778424971332813e-06,
+      "loss": 0.7745,
+      "step": 1900
+    },
+    {
+      "epoch": 0.7353932043507556,
+      "grad_norm": 0.2225854992866516,
+      "learning_rate": 1.7304115736010407e-06,
+      "loss": 0.7731,
+      "step": 1910
+    },
+    {
+      "epoch": 0.739243430551545,
+      "grad_norm": 0.2302289456129074,
+      "learning_rate": 1.6834891968017108e-06,
+      "loss": 0.7654,
+      "step": 1920
+    },
+    {
+      "epoch": 0.7430936567523342,
+      "grad_norm": 0.23412300646305084,
+      "learning_rate": 1.6370826649692884e-06,
+      "loss": 0.7841,
+      "step": 1930
+    },
+    {
+      "epoch": 0.7469438829531235,
+      "grad_norm": 0.2369968146085739,
+      "learning_rate": 1.5911991961040463e-06,
+      "loss": 0.7901,
+      "step": 1940
+    },
+    {
+      "epoch": 0.7507941091539128,
+      "grad_norm": 0.23147568106651306,
+      "learning_rate": 1.545845926849851e-06,
+      "loss": 0.7749,
+      "step": 1950
+    },
+    {
+      "epoch": 0.7546443353547021,
+      "grad_norm": 0.24370233714580536,
+      "learning_rate": 1.5010299113841397e-06,
+      "loss": 0.7945,
+      "step": 1960
+    },
+    {
+      "epoch": 0.7584945615554913,
+      "grad_norm": 0.23152467608451843,
+      "learning_rate": 1.4567581203207238e-06,
+      "loss": 0.7818,
+      "step": 1970
+    },
+    {
+      "epoch": 0.7623447877562807,
+      "grad_norm": 0.2402886152267456,
+      "learning_rate": 1.413037439625593e-06,
+      "loss": 0.7842,
+      "step": 1980
+    },
+    {
+      "epoch": 0.76619501395707,
+      "grad_norm": 0.24002982676029205,
+      "learning_rate": 1.3698746695458847e-06,
+      "loss": 0.7742,
+      "step": 1990
+    },
+    {
+      "epoch": 0.7700452401578592,
+      "grad_norm": 0.22237364947795868,
+      "learning_rate": 1.3272765235521745e-06,
+      "loss": 0.7671,
+      "step": 2000
+    },
+    {
+      "epoch": 0.7738954663586486,
+      "grad_norm": 0.23355218768119812,
+      "learning_rate": 1.2852496272942877e-06,
+      "loss": 0.7757,
+      "step": 2010
+    },
+    {
+      "epoch": 0.7777456925594378,
+      "grad_norm": 0.21445035934448242,
+      "learning_rate": 1.2438005175707396e-06,
+      "loss": 0.7622,
+      "step": 2020
+    },
+    {
+      "epoch": 0.7815959187602272,
+      "grad_norm": 0.2269303798675537,
+      "learning_rate": 1.2029356413120297e-06,
+      "loss": 0.7504,
+      "step": 2030
+    },
+    {
+      "epoch": 0.7854461449610165,
+      "grad_norm": 0.2263043373823166,
+      "learning_rate": 1.1626613545778786e-06,
+      "loss": 0.7653,
+      "step": 2040
+    },
+    {
+      "epoch": 0.7892963711618057,
+      "grad_norm": 0.21538372337818146,
+      "learning_rate": 1.1229839215686294e-06,
+      "loss": 0.7697,
+      "step": 2050
+    },
+    {
+      "epoch": 0.7931465973625951,
+      "grad_norm": 0.20971707999706268,
+      "learning_rate": 1.0839095136509188e-06,
+      "loss": 0.7562,
+      "step": 2060
+    },
+    {
+      "epoch": 0.7969968235633843,
+      "grad_norm": 0.24611474573612213,
+      "learning_rate": 1.045444208397791e-06,
+      "loss": 0.7784,
+      "step": 2070
+    },
+    {
+      "epoch": 0.8008470497641736,
+      "grad_norm": 0.2087804675102234,
+      "learning_rate": 1.007593988643402e-06,
+      "loss": 0.7596,
+      "step": 2080
+    },
+    {
+      "epoch": 0.804697275964963,
+      "grad_norm": 0.21736443042755127,
+      "learning_rate": 9.703647415524613e-07,
+      "loss": 0.7813,
+      "step": 2090
+    },
+    {
+      "epoch": 0.8085475021657522,
+      "grad_norm": 0.22653889656066895,
+      "learning_rate": 9.337622577045485e-07,
+      "loss": 0.7691,
+      "step": 2100
+    },
+    {
+      "epoch": 0.8123977283665416,
+      "grad_norm": 0.22975032031536102,
+      "learning_rate": 8.977922301934588e-07,
+      "loss": 0.7772,
+      "step": 2110
+    },
+    {
+      "epoch": 0.8162479545673308,
+      "grad_norm": 0.23162154853343964,
+      "learning_rate": 8.624602537417009e-07,
+      "loss": 0.7788,
+      "step": 2120
+    },
+    {
+      "epoch": 0.8200981807681201,
+      "grad_norm": 0.24134595692157745,
+      "learning_rate": 8.277718238303123e-07,
+      "loss": 0.7656,
+      "step": 2130
+    },
+    {
+      "epoch": 0.8239484069689095,
+      "grad_norm": 0.22465036809444427,
+      "learning_rate": 7.937323358440935e-07,
+      "loss": 0.7697,
+      "step": 2140
+    },
+    {
+      "epoch": 0.8277986331696987,
+      "grad_norm": 0.24127823114395142,
+      "learning_rate": 7.603470842324223e-07,
+      "loss": 0.7616,
+      "step": 2150
+    },
+    {
+      "epoch": 0.831648859370488,
+      "grad_norm": 0.22312285006046295,
+      "learning_rate": 7.276212616857575e-07,
+      "loss": 0.78,
+      "step": 2160
+    },
+    {
+      "epoch": 0.8354990855712773,
+      "grad_norm": 0.2267819344997406,
+      "learning_rate": 6.955599583279837e-07,
+      "loss": 0.7814,
+      "step": 2170
+    },
+    {
+      "epoch": 0.8393493117720666,
+      "grad_norm": 0.21880404651165009,
+      "learning_rate": 6.641681609246981e-07,
+      "loss": 0.7712,
+      "step": 2180
+    },
+    {
+      "epoch": 0.8431995379728559,
+      "grad_norm": 0.21805761754512787,
+      "learning_rate": 6.334507521075711e-07,
+      "loss": 0.7808,
+      "step": 2190
+    },
+    {
+      "epoch": 0.8470497641736452,
+      "grad_norm": 0.24373184144496918,
+      "learning_rate": 6.034125096149179e-07,
+      "loss": 0.7742,
+      "step": 2200
+    },
+    {
+      "epoch": 0.8508999903744345,
+      "grad_norm": 0.2268376350402832,
+      "learning_rate": 5.740581055485722e-07,
+      "loss": 0.7438,
+      "step": 2210
+    },
+    {
+      "epoch": 0.8547502165752238,
+      "grad_norm": 0.2106325477361679,
+      "learning_rate": 5.453921056471961e-07,
+      "loss": 0.7771,
+      "step": 2220
+    },
+    {
+      "epoch": 0.8586004427760131,
+      "grad_norm": 0.22670277953147888,
+      "learning_rate": 5.174189685761288e-07,
+      "loss": 0.7719,
+      "step": 2230
+    },
+    {
+      "epoch": 0.8624506689768023,
+      "grad_norm": 0.21321237087249756,
+      "learning_rate": 4.901430452338979e-07,
+      "loss": 0.7829,
+      "step": 2240
+    },
+    {
+      "epoch": 0.8663008951775917,
+      "grad_norm": 0.22154928743839264,
+      "learning_rate": 4.6356857807548286e-07,
+      "loss": 0.769,
+      "step": 2250
+    },
+    {
+      "epoch": 0.870151121378381,
+      "grad_norm": 0.22588704526424408,
+      "learning_rate": 4.376997004524541e-07,
+      "loss": 0.7585,
+      "step": 2260
+    },
+    {
+      "epoch": 0.8740013475791703,
+      "grad_norm": 0.22505638003349304,
+      "learning_rate": 4.12540435970073e-07,
+      "loss": 0.767,
+      "step": 2270
+    },
+    {
+      "epoch": 0.8778515737799596,
+      "grad_norm": 0.2229229360818863,
+      "learning_rate": 3.880946978614703e-07,
+      "loss": 0.7747,
+      "step": 2280
+    },
+    {
+      "epoch": 0.8817017999807488,
+      "grad_norm": 0.22818392515182495,
+      "learning_rate": 3.643662883789878e-07,
+      "loss": 0.7815,
+      "step": 2290
+    },
+    {
+      "epoch": 0.8855520261815382,
+      "grad_norm": 0.2079848349094391,
+      "learning_rate": 3.41358898202781e-07,
+      "loss": 0.7601,
+      "step": 2300
+    },
+    {
+      "epoch": 0.8894022523823275,
+      "grad_norm": 0.20991580188274384,
+      "learning_rate": 3.190761058667741e-07,
+      "loss": 0.7727,
+      "step": 2310
+    },
+    {
+      "epoch": 0.8932524785831167,
+      "grad_norm": 0.22084525227546692,
+      "learning_rate": 2.9752137720206463e-07,
+      "loss": 0.7761,
+      "step": 2320
+    },
+    {
+      "epoch": 0.8971027047839061,
+      "grad_norm": 0.2241315096616745,
+      "learning_rate": 2.7669806479785067e-07,
+      "loss": 0.7808,
+      "step": 2330
+    },
+    {
+      "epoch": 0.9009529309846953,
+      "grad_norm": 0.22336702048778534,
+      "learning_rate": 2.566094074799763e-07,
+      "loss": 0.7761,
+      "step": 2340
+    },
+    {
+      "epoch": 0.9048031571854847,
+      "grad_norm": 0.2216758131980896,
+      "learning_rate": 2.3725852980716557e-07,
+      "loss": 0.7582,
+      "step": 2350
+    },
+    {
+      "epoch": 0.908653383386274,
+      "grad_norm": 0.2207355797290802,
+      "learning_rate": 2.1864844158503983e-07,
+      "loss": 0.7705,
+      "step": 2360
+    },
+    {
+      "epoch": 0.9125036095870632,
+      "grad_norm": 0.24372541904449463,
+      "learning_rate": 2.007820373979702e-07,
+      "loss": 0.762,
+      "step": 2370
+    },
+    {
+      "epoch": 0.9163538357878526,
+      "grad_norm": 0.23084378242492676,
+      "learning_rate": 1.8366209615886532e-07,
+      "loss": 0.7725,
+      "step": 2380
+    },
+    {
+      "epoch": 0.9202040619886418,
+      "grad_norm": 0.22515594959259033,
+      "learning_rate": 1.6729128067693613e-07,
+      "loss": 0.78,
+      "step": 2390
+    },
+    {
+      "epoch": 0.9240542881894311,
+      "grad_norm": 0.2390122264623642,
+      "learning_rate": 1.5167213724353426e-07,
+      "loss": 0.7813,
+      "step": 2400
+    },
+    {
+      "epoch": 0.9279045143902205,
+      "grad_norm": 0.22326770424842834,
+      "learning_rate": 1.3680709523610058e-07,
+      "loss": 0.7611,
+      "step": 2410
+    },
+    {
+      "epoch": 0.9317547405910097,
+      "grad_norm": 0.2288760244846344,
+      "learning_rate": 1.2269846674030418e-07,
+      "loss": 0.7874,
+      "step": 2420
+    },
+    {
+      "epoch": 0.935604966791799,
+      "grad_norm": 0.23877480626106262,
+      "learning_rate": 1.0934844619042595e-07,
+      "loss": 0.781,
+      "step": 2430
+    },
+    {
+      "epoch": 0.9394551929925883,
+      "grad_norm": 0.23175378143787384,
+      "learning_rate": 9.675911002803961e-08,
+      "loss": 0.7712,
+      "step": 2440
+    },
+    {
+      "epoch": 0.9433054191933776,
+      "grad_norm": 0.23841503262519836,
+      "learning_rate": 8.493241637904381e-08,
+      "loss": 0.7699,
+      "step": 2450
+    },
+    {
+      "epoch": 0.9471556453941669,
+      "grad_norm": 0.22205816209316254,
+      "learning_rate": 7.387020474909689e-08,
+      "loss": 0.7867,
+      "step": 2460
+    },
+    {
+      "epoch": 0.9510058715949562,
+      "grad_norm": 0.24204526841640472,
+      "learning_rate": 6.357419573750856e-08,
+      "loss": 0.7843,
+      "step": 2470
+    },
+    {
+      "epoch": 0.9548560977957455,
+      "grad_norm": 0.22353723645210266,
+      "learning_rate": 5.4045990769613365e-08,
+      "loss": 0.7651,
+      "step": 2480
+    },
+    {
+      "epoch": 0.9587063239965348,
+      "grad_norm": 0.22606970369815826,
+      "learning_rate": 4.5287071847694434e-08,
+      "loss": 0.773,
+      "step": 2490
+    },
+    {
+      "epoch": 0.9625565501973241,
+      "grad_norm": 0.22452645003795624,
+      "learning_rate": 3.7298801320467817e-08,
+      "loss": 0.7669,
+      "step": 2500
+    },
+    {
+      "epoch": 0.9664067763981133,
+      "grad_norm": 0.22227899730205536,
+      "learning_rate": 3.0082421671192576e-08,
+      "loss": 0.76,
+      "step": 2510
+    },
+    {
+      "epoch": 0.9702570025989027,
+      "grad_norm": 0.2222938984632492,
+      "learning_rate": 2.3639055324410355e-08,
+      "loss": 0.7788,
+      "step": 2520
+    },
+    {
+      "epoch": 0.974107228799692,
+      "grad_norm": 0.22521865367889404,
+      "learning_rate": 1.7969704471369477e-08,
+      "loss": 0.7666,
+      "step": 2530
+    },
+    {
+      "epoch": 0.9779574550004813,
+      "grad_norm": 0.2158743143081665,
+      "learning_rate": 1.3075250914142413e-08,
+      "loss": 0.7675,
+      "step": 2540
+    },
+    {
+      "epoch": 0.9818076812012706,
+      "grad_norm": 0.21967364847660065,
+      "learning_rate": 8.956455928476049e-09,
+      "loss": 0.7782,
+      "step": 2550
+    },
+    {
+      "epoch": 0.9856579074020598,
+      "grad_norm": 0.22446706891059875,
+      "learning_rate": 5.6139601453786275e-09,
+      "loss": 0.7711,
+      "step": 2560
+    },
+    {
+      "epoch": 0.9895081336028492,
+      "grad_norm": 0.2197936624288559,
+      "learning_rate": 3.0482834514805602e-09,
+      "loss": 0.7605,
+      "step": 2570
+    },
+    {
+      "epoch": 0.9933583598036385,
+      "grad_norm": 0.21729329228401184,
+      "learning_rate": 1.2598249081713366e-09,
+      "loss": 0.753,
+      "step": 2580
+    },
+    {
+      "epoch": 0.9972085860044277,
+      "grad_norm": 0.23097607493400574,
+      "learning_rate": 2.488626895302826e-10,
+      "loss": 0.7837,
+      "step": 2590
+    },
+    {
+      "epoch": 0.9999037443449803,
+      "step": 2597,
+      "total_flos": 3.517726782637736e+18,
+      "train_loss": 0.7911533439072178,
+      "train_runtime": 10002.5075,
+      "train_samples_per_second": 66.474,
+      "train_steps_per_second": 0.26
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 2597,
+  "num_input_tokens_seen": 418855352,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 3.517726782637736e+18,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

continuous/qwen25/pixio_vitb16/finetune/training_args.bin ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:cbd826095d0ba0cccf66ba20fa00ee03c8064bb886341025bb26cf6b779838b5
+size 6264

continuous/qwen25/pixio_vitb16/finetune/vocab.json ADDED Viewed

The diff for this file is too large to render. See raw diff

continuous/qwen25/pixio_vitb16/pretrain/config.json ADDED Viewed

	@@ -0,0 +1,87 @@

+{
+  "add_faster_video": false,
+  "add_time_instruction": false,
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "dtype": "bfloat16",
+  "eos_token_id": 151645,
+  "faster_token_stride": 10,
+  "force_sample": false,
+  "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,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "mm_hidden_size": 768,
+  "mm_newline_position": "grid",
+  "mm_patch_merge_type": "flat",
+  "mm_projector_lr": null,
+  "mm_projector_type": "mlp2x_gelu",
+  "mm_resampler_type": null,
+  "mm_spatial_pool_mode": "bilinear",
+  "mm_spatial_pool_stride": null,
+  "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": -1,
+  "mm_vision_tower": "vtb_ssl:pixio",
+  "mm_vision_tower_lr": null,
+  "model_type": "qwen2",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 2,
+  "pad_token_id": 151643,
+  "pos_skipping_range": 4096,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "tokenizer_model_max_length": 8192,
+  "tokenizer_padding_side": "right",
+  "transformers_version": "4.57.6",
+  "use_cache": true,
+  "use_mm_proj": true,
+  "use_pos_skipping": false,
+  "use_sliding_window": false,
+  "vision_tower_pretrained": "/cache/ckpt/download/tokenizer/continuous/pixio/pixio-vitb16/pixio_vitb16.pth",
+  "vocab_size": 151936
+}

continuous/qwen25/pixio_vitb16/pretrain/mm_projector.bin ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:754ca4a514f244bcc41d486b7706328942de20800895829682858dad6fed16e5
+size 14170168

continuous/qwen25/pixio_vitb16/pretrain/trainer_state.json ADDED Viewed

	@@ -0,0 +1,421 @@

+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 545,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01834862385321101,
+      "grad_norm": 0.18827709555625916,
+      "learning_rate": 0.00032142857142857147,
+      "loss": 4.7681,
+      "step": 10
+    },
+    {
+      "epoch": 0.03669724770642202,
+      "grad_norm": 0.06294339895248413,
+      "learning_rate": 0.0006785714285714287,
+      "loss": 3.8888,
+      "step": 20
+    },
+    {
+      "epoch": 0.05504587155963303,
+      "grad_norm": 0.029645342379808426,
+      "learning_rate": 0.0009999907688176173,
+      "loss": 3.4666,
+      "step": 30
+    },
+    {
+      "epoch": 0.07339449541284404,
+      "grad_norm": 0.024336915463209152,
+      "learning_rate": 0.0009988834393115766,
+      "loss": 3.3299,
+      "step": 40
+    },
+    {
+      "epoch": 0.09174311926605505,
+      "grad_norm": 0.037347763776779175,
+      "learning_rate": 0.0009959345572940772,
+      "loss": 3.282,
+      "step": 50
+    },
+    {
+      "epoch": 0.11009174311926606,
+      "grad_norm": 0.017387213185429573,
+      "learning_rate": 0.0009911550081155983,
+      "loss": 3.2502,
+      "step": 60
+    },
+    {
+      "epoch": 0.12844036697247707,
+      "grad_norm": 0.018883930519223213,
+      "learning_rate": 0.0009845624347566433,
+      "loss": 3.2192,
+      "step": 70
+    },
+    {
+      "epoch": 0.14678899082568808,
+      "grad_norm": 0.03376435860991478,
+      "learning_rate": 0.0009761811727013547,
+      "loss": 3.1928,
+      "step": 80
+    },
+    {
+      "epoch": 0.1651376146788991,
+      "grad_norm": 0.0194522924721241,
+      "learning_rate": 0.0009660421601067666,
+      "loss": 3.1625,
+      "step": 90
+    },
+    {
+      "epoch": 0.1834862385321101,
+      "grad_norm": 0.0338672399520874,
+      "learning_rate": 0.0009541828235992926,
+      "loss": 3.1452,
+      "step": 100
+    },
+    {
+      "epoch": 0.2018348623853211,
+      "grad_norm": 0.0240735225379467,
+      "learning_rate": 0.0009406469401200151,
+      "loss": 3.1159,
+      "step": 110
+    },
+    {
+      "epoch": 0.22018348623853212,
+      "grad_norm": 0.03028913214802742,
+      "learning_rate": 0.0009254844753287492,
+      "loss": 3.1026,
+      "step": 120
+    },
+    {
+      "epoch": 0.23853211009174313,
+      "grad_norm": 0.03244874253869057,
+      "learning_rate": 0.0009087513991633924,
+      "loss": 3.0682,
+      "step": 130
+    },
+    {
+      "epoch": 0.25688073394495414,
+      "grad_norm": 0.025908568874001503,
+      "learning_rate": 0.0008905094792353917,
+      "loss": 3.0512,
+      "step": 140
+    },
+    {
+      "epoch": 0.27522935779816515,
+      "grad_norm": 0.028307348489761353,
+      "learning_rate": 0.0008708260528239789,
+      "loss": 3.0409,
+      "step": 150
+    },
+    {
+      "epoch": 0.29357798165137616,
+      "grad_norm": 0.026526091620326042,
+      "learning_rate": 0.0008497737783108237,
+      "loss": 3.0244,
+      "step": 160
+    },
+    {
+      "epoch": 0.3119266055045872,
+      "grad_norm": 0.022967547178268433,
+      "learning_rate": 0.0008274303669726426,
+      "loss": 3.0152,
+      "step": 170
+    },
+    {
+      "epoch": 0.3302752293577982,
+      "grad_norm": 0.02606263756752014,
+      "learning_rate": 0.0008038782961218137,
+      "loss": 2.9974,
+      "step": 180
+    },
+    {
+      "epoch": 0.3486238532110092,
+      "grad_norm": 0.02934882789850235,
+      "learning_rate": 0.000779204504653894,
+      "loss": 2.993,
+      "step": 190
+    },
+    {
+      "epoch": 0.3669724770642202,
+      "grad_norm": 0.023173265159130096,
+      "learning_rate": 0.0007535000721258791,
+      "loss": 2.9719,
+      "step": 200
+    },
+    {
+      "epoch": 0.3853211009174312,
+      "grad_norm": 0.03466491773724556,
+      "learning_rate": 0.00072685988254984,
+      "loss": 2.9683,
+      "step": 210
+    },
+    {
+      "epoch": 0.4036697247706422,
+      "grad_norm": 0.024940554052591324,
+      "learning_rate": 0.0006993822741429906,
+      "loss": 2.9779,
+      "step": 220
+    },
+    {
+      "epoch": 0.42201834862385323,
+      "grad_norm": 0.022818198427557945,
+      "learning_rate": 0.0006711686763270818,
+      "loss": 2.9696,
+      "step": 230
+    },
+    {
+      "epoch": 0.44036697247706424,
+      "grad_norm": 0.023421572521328926,
+      "learning_rate": 0.0006423232353170831,
+      "loss": 2.9554,
+      "step": 240
+    },
+    {
+      "epoch": 0.45871559633027525,
+      "grad_norm": 0.022569596767425537,
+      "learning_rate": 0.0006129524296812335,
+      "loss": 2.9418,
+      "step": 250
+    },
+    {
+      "epoch": 0.47706422018348627,
+      "grad_norm": 0.022295331582427025,
+      "learning_rate": 0.0005831646772915651,
+      "loss": 2.9542,
+      "step": 260
+    },
+    {
+      "epoch": 0.4954128440366973,
+      "grad_norm": 0.023684734478592873,
+      "learning_rate": 0.0005530699351157809,
+      "loss": 2.9346,
+      "step": 270
+    },
+    {
+      "epoch": 0.5137614678899083,
+      "grad_norm": 0.021011192351579666,
+      "learning_rate": 0.0005227792933277943,
+      "loss": 2.9363,
+      "step": 280
+    },
+    {
+      "epoch": 0.5321100917431193,
+      "grad_norm": 0.021929055452346802,
+      "learning_rate": 0.0004924045652352092,
+      "loss": 2.9172,
+      "step": 290
+    },
+    {
+      "epoch": 0.5504587155963303,
+      "grad_norm": 0.019937578588724136,
+      "learning_rate": 0.0004620578745374633,
+      "loss": 2.9175,
+      "step": 300
+    },
+    {
+      "epoch": 0.5688073394495413,
+      "grad_norm": 0.025011789053678513,
+      "learning_rate": 0.0004318512414382059,
+      "loss": 2.9258,
+      "step": 310
+    },
+    {
+      "epoch": 0.5871559633027523,
+      "grad_norm": 0.020348789170384407,
+      "learning_rate": 0.0004018961691397148,
+      "loss": 2.9216,
+      "step": 320
+    },
+    {
+      "epoch": 0.6055045871559633,
+      "grad_norm": 0.024784652516245842,
+      "learning_rate": 0.0003723032322457458,
+      "loss": 2.9071,
+      "step": 330
+    },
+    {
+      "epoch": 0.6238532110091743,
+      "grad_norm": 0.03244921565055847,
+      "learning_rate": 0.0003431816685921625,
+      "loss": 2.9108,
+      "step": 340
+    },
+    {
+      "epoch": 0.6422018348623854,
+      "grad_norm": 0.020494218915700912,
+      "learning_rate": 0.0003146389760120416,
+      "loss": 2.9138,
+      "step": 350
+    },
+    {
+      "epoch": 0.6605504587155964,
+      "grad_norm": 0.026060283184051514,
+      "learning_rate": 0.00028678051552373485,
+      "loss": 2.9131,
+      "step": 360
+    },
+    {
+      "epoch": 0.6788990825688074,
+      "grad_norm": 0.022106077522039413,
+      "learning_rate": 0.0002597091224066581,
+      "loss": 2.9024,
+      "step": 370
+    },
+    {
+      "epoch": 0.6972477064220184,
+      "grad_norm": 0.018173808231949806,
+      "learning_rate": 0.0002335247266004629,
+      "loss": 2.9158,
+      "step": 380
+    },
+    {
+      "epoch": 0.7155963302752294,
+      "grad_norm": 0.022986184805631638,
+      "learning_rate": 0.000208323983828828,
+      "loss": 2.9014,
+      "step": 390
+    },
+    {
+      "epoch": 0.7339449541284404,
+      "grad_norm": 0.018242252990603447,
+      "learning_rate": 0.0001841999188095224,
+      "loss": 2.8964,
+      "step": 400
+    },
+    {
+      "epoch": 0.7522935779816514,
+      "grad_norm": 0.018277421593666077,
+      "learning_rate": 0.00016124158186777676,
+      "loss": 2.8957,
+      "step": 410
+    },
+    {
+      "epoch": 0.7706422018348624,
+      "grad_norm": 0.021453872323036194,
+      "learning_rate": 0.0001395337202205211,
+      "loss": 2.8892,
+      "step": 420
+    },
+    {
+      "epoch": 0.7889908256880734,
+      "grad_norm": 0.018292300403118134,
+      "learning_rate": 0.00011915646514489292,
+      "loss": 2.8925,
+      "step": 430
+    },
+    {
+      "epoch": 0.8073394495412844,
+      "grad_norm": 0.021583616733551025,
+      "learning_rate": 0.00010018503618578818,
+      "loss": 2.9012,
+      "step": 440
+    },
+    {
+      "epoch": 0.8256880733944955,
+      "grad_norm": 0.01806589402258396,
+      "learning_rate": 8.268946349432583e-05,
+      "loss": 2.8965,
+      "step": 450
+    },
+    {
+      "epoch": 0.8440366972477065,
+      "grad_norm": 0.017329780384898186,
+      "learning_rate": 6.673432932217117e-05,
+      "loss": 2.8849,
+      "step": 460
+    },
+    {
+      "epoch": 0.8623853211009175,
+      "grad_norm": 0.016085077077150345,
+      "learning_rate": 5.2378529625954684e-05,
+      "loss": 2.8764,
+      "step": 470
+    },
+    {
+      "epoch": 0.8807339449541285,
+      "grad_norm": 0.01703362911939621,
+      "learning_rate": 3.9675056661785556e-05,
+      "loss": 2.8844,
+      "step": 480
+    },
+    {
+      "epoch": 0.8990825688073395,
+      "grad_norm": 0.018842943012714386,
+      "learning_rate": 2.8670803372379538e-05,
+      "loss": 2.8889,
+      "step": 490
+    },
+    {
+      "epoch": 0.9174311926605505,
+      "grad_norm": 0.01654229499399662,
+      "learning_rate": 1.9406390288876586e-05,
+      "loss": 2.8985,
+      "step": 500
+    },
+    {
+      "epoch": 0.9357798165137615,
+      "grad_norm": 0.017145950347185135,
+      "learning_rate": 1.191601558631461e-05,
+      "loss": 2.8852,
+      "step": 510
+    },
+    {
+      "epoch": 0.9541284403669725,
+      "grad_norm": 0.016451647505164146,
+      "learning_rate": 6.2273288462559305e-06,
+      "loss": 2.8814,
+      "step": 520
+    },
+    {
+      "epoch": 0.9724770642201835,
+      "grad_norm": 0.01736357808113098,
+      "learning_rate": 2.361328992552314e-06,
+      "loss": 2.8819,
+      "step": 530
+    },
+    {
+      "epoch": 0.9908256880733946,
+      "grad_norm": 0.016590874642133713,
+      "learning_rate": 3.322867770044202e-07,
+      "loss": 2.8845,
+      "step": 540
+    },
+    {
+      "epoch": 1.0,
+      "step": 545,
+      "total_flos": 2.4527281736332083e+17,
+      "train_loss": 3.0377867234956235,
+      "train_runtime": 5261.2791,
+      "train_samples_per_second": 106.082,
+      "train_steps_per_second": 0.104
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 545,
+  "num_input_tokens_seen": 29204608,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": false,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.4527281736332083e+17,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/added_tokens.json ADDED Viewed

	@@ -0,0 +1,24 @@

+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/chat_template.jinja ADDED Viewed

	@@ -0,0 +1,54 @@

+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- messages[0]['content'] }}
+    {%- else %}
+        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
+    {%- endif %}
+    {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
+    {%- else %}
+        {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role }}
+        {%- if message.content %}
+            {{- '\n' + message.content }}
+        {%- endif %}
+        {%- for tool_call in message.tool_calls %}
+            {%- if tool_call.function is defined %}
+                {%- set tool_call = tool_call.function %}
+            {%- endif %}
+            {{- '\n<tool_call>\n{"name": "' }}
+            {{- tool_call.name }}
+            {{- '", "arguments": ' }}
+            {{- tool_call.arguments | tojson }}
+            {{- '}\n</tool_call>' }}
+        {%- endfor %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+{%- endif %}

continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/config.json ADDED Viewed

	@@ -0,0 +1,87 @@

+{
+  "add_faster_video": false,
+  "add_time_instruction": false,
+  "architectures": [
+    "LlavaQwenForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "dtype": "bfloat16",
+  "eos_token_id": 151645,
+  "faster_token_stride": 10,
+  "force_sample": false,
+  "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,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "mm_hidden_size": 1280,
+  "mm_newline_position": "grid",
+  "mm_patch_merge_type": "flat",
+  "mm_projector_lr": 0.0001,
+  "mm_projector_type": "mlp2x_gelu",
+  "mm_resampler_type": null,
+  "mm_spatial_pool_mode": "bilinear",
+  "mm_spatial_pool_stride": null,
+  "mm_tunable_parts": "mm_mlp_adapter,mm_language_model",
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -1,
+  "mm_vision_tower": "vtb_ssl:pixio",
+  "mm_vision_tower_lr": null,
+  "model_type": "qwen2",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 2,
+  "pad_token_id": 151643,
+  "pos_skipping_range": 4096,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "tokenizer_model_max_length": 8192,
+  "tokenizer_padding_side": "right",
+  "transformers_version": "4.57.6",
+  "use_cache": true,
+  "use_mm_proj": true,
+  "use_pos_skipping": false,
+  "use_sliding_window": false,
+  "vision_tower_pretrained": "/cache/ckpt/download/tokenizer/continuous/pixio/pixio_vith16/model.native.pth",
+  "vocab_size": 151936
+}

continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/generation_config.json ADDED Viewed

	@@ -0,0 +1,13 @@

+{
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.1,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "4.57.6"
+}

continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/merges.txt ADDED Viewed

The diff for this file is too large to render. See raw diff

continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/model.safetensors ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:7206f37a7d5a96d50a749662d49f24703b1ac9a4cf5ded7479b521355820b1f7
+size 4833605896

continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/special_tokens_map.json ADDED Viewed

	@@ -0,0 +1,31 @@

+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/tokenizer.json ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:9c5ae00e602b8860cbd784ba82a8aa14e8feecec692e7076590d014d7b7fdafa
+size 11421896

continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/tokenizer_config.json ADDED Viewed

	@@ -0,0 +1,208 @@

+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 8192,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/trainer_state.json ADDED Viewed

	@@ -0,0 +1,1856 @@

+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9999037443449803,
+  "eval_steps": 500,
+  "global_step": 2597,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.003850226200789296,
+      "grad_norm": 4.5110859870910645,
+      "learning_rate": 1.153846153846154e-06,
+      "loss": 2.084,
+      "step": 10
+    },
+    {
+      "epoch": 0.007700452401578592,
+      "grad_norm": 3.996511936187744,
+      "learning_rate": 2.435897435897436e-06,
+      "loss": 1.7377,
+      "step": 20
+    },
+    {
+      "epoch": 0.01155067860236789,
+      "grad_norm": 1.2254811525344849,
+      "learning_rate": 3.7179487179487184e-06,
+      "loss": 1.2014,
+      "step": 30
+    },
+    {
+      "epoch": 0.015400904803157185,
+      "grad_norm": 0.26869136095046997,
+      "learning_rate": 5e-06,
+      "loss": 0.9901,
+      "step": 40
+    },
+    {
+      "epoch": 0.019251131003946483,
+      "grad_norm": 0.1428574025630951,
+      "learning_rate": 6.282051282051282e-06,
+      "loss": 0.9541,
+      "step": 50
+    },
+    {
+      "epoch": 0.02310135720473578,
+      "grad_norm": 0.1062077209353447,
+      "learning_rate": 7.564102564102564e-06,
+      "loss": 0.9055,
+      "step": 60
+    },
+    {
+      "epoch": 0.026951583405525073,
+      "grad_norm": 0.09064631909132004,
+      "learning_rate": 8.846153846153847e-06,
+      "loss": 0.8784,
+      "step": 70
+    },
+    {
+      "epoch": 0.03080180960631437,
+      "grad_norm": 0.08711939305067062,
+      "learning_rate": 9.999996111488723e-06,
+      "loss": 0.8941,
+      "step": 80
+    },
+    {
+      "epoch": 0.034652035807103666,
+      "grad_norm": 0.07945314794778824,
+      "learning_rate": 9.999529497453782e-06,
+      "loss": 0.8676,
+      "step": 90
+    },
+    {
+      "epoch": 0.038502262007892966,
+      "grad_norm": 0.0693436861038208,
+      "learning_rate": 9.998285264324193e-06,
+      "loss": 0.8463,
+      "step": 100
+    },
+    {
+      "epoch": 0.04235248820868226,
+      "grad_norm": 0.06705448031425476,
+      "learning_rate": 9.996263605626052e-06,
+      "loss": 0.8479,
+      "step": 110
+    },
+    {
+      "epoch": 0.04620271440947156,
+      "grad_norm": 0.07512293756008148,
+      "learning_rate": 9.99346483580503e-06,
+      "loss": 0.8417,
+      "step": 120
+    },
+    {
+      "epoch": 0.05005294061026085,
+      "grad_norm": 0.06506039947271347,
+      "learning_rate": 9.989889390177467e-06,
+      "loss": 0.8267,
+      "step": 130
+    },
+    {
+      "epoch": 0.053903166811050146,
+      "grad_norm": 0.06580442190170288,
+      "learning_rate": 9.985537824862647e-06,
+      "loss": 0.8324,
+      "step": 140
+    },
+    {
+      "epoch": 0.057753393011839446,
+      "grad_norm": 0.07490972429513931,
+      "learning_rate": 9.980410816696317e-06,
+      "loss": 0.8337,
+      "step": 150
+    },
+    {
+      "epoch": 0.06160361921262874,
+      "grad_norm": 0.06513803452253342,
+      "learning_rate": 9.97450916312541e-06,
+      "loss": 0.8321,
+      "step": 160
+    },
+    {
+      "epoch": 0.06545384541341803,
+      "grad_norm": 0.05901731923222542,
+      "learning_rate": 9.967833782084004e-06,
+      "loss": 0.8126,
+      "step": 170
+    },
+    {
+      "epoch": 0.06930407161420733,
+      "grad_norm": 0.06660837680101395,
+      "learning_rate": 9.960385711850554e-06,
+      "loss": 0.847,
+      "step": 180
+    },
+    {
+      "epoch": 0.07315429781499663,
+      "grad_norm": 0.0666705071926117,
+      "learning_rate": 9.952166110886394e-06,
+      "loss": 0.8458,
+      "step": 190
+    },
+    {
+      "epoch": 0.07700452401578593,
+      "grad_norm": 0.07041464000940323,
+      "learning_rate": 9.943176257655567e-06,
+      "loss": 0.8191,
+      "step": 200
+    },
+    {
+      "epoch": 0.08085475021657522,
+      "grad_norm": 0.06317471712827682,
+      "learning_rate": 9.933417550425952e-06,
+      "loss": 0.8274,
+      "step": 210
+    },
+    {
+      "epoch": 0.08470497641736452,
+      "grad_norm": 0.06406091898679733,
+      "learning_rate": 9.9228915070518e-06,
+      "loss": 0.8154,
+      "step": 220
+    },
+    {
+      "epoch": 0.08855520261815382,
+      "grad_norm": 0.06682354211807251,
+      "learning_rate": 9.911599764737632e-06,
+      "loss": 0.8091,
+      "step": 230
+    },
+    {
+      "epoch": 0.09240542881894312,
+      "grad_norm": 0.07348091900348663,
+      "learning_rate": 9.899544079783606e-06,
+      "loss": 0.8059,
+      "step": 240
+    },
+    {
+      "epoch": 0.0962556550197324,
+      "grad_norm": 0.06326447427272797,
+      "learning_rate": 9.886726327312338e-06,
+      "loss": 0.8095,
+      "step": 250
+    },
+    {
+      "epoch": 0.1001058812205217,
+      "grad_norm": 0.06674189120531082,
+      "learning_rate": 9.87314850097724e-06,
+      "loss": 0.8179,
+      "step": 260
+    },
+    {
+      "epoch": 0.103956107421311,
+      "grad_norm": 0.06721276789903641,
+      "learning_rate": 9.858812712652452e-06,
+      "loss": 0.8165,
+      "step": 270
+    },
+    {
+      "epoch": 0.10780633362210029,
+      "grad_norm": 0.059832558035850525,
+      "learning_rate": 9.843721192104341e-06,
+      "loss": 0.8089,
+      "step": 280
+    },
+    {
+      "epoch": 0.11165655982288959,
+      "grad_norm": 0.06226849555969238,
+      "learning_rate": 9.827876286644698e-06,
+      "loss": 0.8226,
+      "step": 290
+    },
+    {
+      "epoch": 0.11550678602367889,
+      "grad_norm": 0.06116672232747078,
+      "learning_rate": 9.811280460765643e-06,
+      "loss": 0.7919,
+      "step": 300
+    },
+    {
+      "epoch": 0.11935701222446819,
+      "grad_norm": 0.06354665011167526,
+      "learning_rate": 9.793936295756292e-06,
+      "loss": 0.8026,
+      "step": 310
+    },
+    {
+      "epoch": 0.12320723842525748,
+      "grad_norm": 0.06112486496567726,
+      "learning_rate": 9.775846489301279e-06,
+      "loss": 0.8014,
+      "step": 320
+    },
+    {
+      "epoch": 0.12705746462604678,
+      "grad_norm": 0.06482264399528503,
+      "learning_rate": 9.757013855061157e-06,
+      "loss": 0.8045,
+      "step": 330
+    },
+    {
+      "epoch": 0.13090769082683607,
+      "grad_norm": 0.06127361208200455,
+      "learning_rate": 9.737441322234758e-06,
+      "loss": 0.8128,
+      "step": 340
+    },
+    {
+      "epoch": 0.13475791702762538,
+      "grad_norm": 0.06512747704982758,
+      "learning_rate": 9.717131935103605e-06,
+      "loss": 0.8118,
+      "step": 350
+    },
+    {
+      "epoch": 0.13860814322841467,
+      "grad_norm": 0.057247359305620193,
+      "learning_rate": 9.696088852558398e-06,
+      "loss": 0.7971,
+      "step": 360
+    },
+    {
+      "epoch": 0.14245836942920398,
+      "grad_norm": 0.06131359189748764,
+      "learning_rate": 9.674315347607686e-06,
+      "loss": 0.8315,
+      "step": 370
+    },
+    {
+      "epoch": 0.14630859562999327,
+      "grad_norm": 0.06116604804992676,
+      "learning_rate": 9.651814806868796e-06,
+      "loss": 0.7888,
+      "step": 380
+    },
+    {
+      "epoch": 0.15015882183078255,
+      "grad_norm": 0.06148853898048401,
+      "learning_rate": 9.62859073004107e-06,
+      "loss": 0.802,
+      "step": 390
+    },
+    {
+      "epoch": 0.15400904803157187,
+      "grad_norm": 0.05753828585147858,
+      "learning_rate": 9.604646729361547e-06,
+      "loss": 0.7991,
+      "step": 400
+    },
+    {
+      "epoch": 0.15785927423236115,
+      "grad_norm": 0.06215183436870575,
+      "learning_rate": 9.579986529043099e-06,
+      "loss": 0.796,
+      "step": 410
+    },
+    {
+      "epoch": 0.16170950043315044,
+      "grad_norm": 0.06532777100801468,
+      "learning_rate": 9.554613964695189e-06,
+      "loss": 0.7919,
+      "step": 420
+    },
+    {
+      "epoch": 0.16555972663393975,
+      "grad_norm": 0.06405987590551376,
+      "learning_rate": 9.528532982727281e-06,
+      "loss": 0.7919,
+      "step": 430
+    },
+    {
+      "epoch": 0.16940995283472904,
+      "grad_norm": 0.07137701660394669,
+      "learning_rate": 9.501747639735018e-06,
+      "loss": 0.8057,
+      "step": 440
+    },
+    {
+      "epoch": 0.17326017903551832,
+      "grad_norm": 0.06073567643761635,
+      "learning_rate": 9.474262101869272e-06,
+      "loss": 0.7776,
+      "step": 450
+    },
+    {
+      "epoch": 0.17711040523630764,
+      "grad_norm": 0.05947998911142349,
+      "learning_rate": 9.446080644188141e-06,
+      "loss": 0.7951,
+      "step": 460
+    },
+    {
+      "epoch": 0.18096063143709692,
+      "grad_norm": 0.06650283932685852,
+      "learning_rate": 9.417207649992017e-06,
+      "loss": 0.8091,
+      "step": 470
+    },
+    {
+      "epoch": 0.18481085763788624,
+      "grad_norm": 0.05980244278907776,
+      "learning_rate": 9.387647610141816e-06,
+      "loss": 0.7914,
+      "step": 480
+    },
+    {
+      "epoch": 0.18866108383867553,
+      "grad_norm": 0.055913250893354416,
+      "learning_rate": 9.35740512236047e-06,
+      "loss": 0.793,
+      "step": 490
+    },
+    {
+      "epoch": 0.1925113100394648,
+      "grad_norm": 0.05581941083073616,
+      "learning_rate": 9.326484890517809e-06,
+      "loss": 0.7918,
+      "step": 500
+    },
+    {
+      "epoch": 0.19636153624025413,
+      "grad_norm": 0.06568170338869095,
+      "learning_rate": 9.294891723898926e-06,
+      "loss": 0.7694,
+      "step": 510
+    },
+    {
+      "epoch": 0.2002117624410434,
+      "grad_norm": 0.07223556190729141,
+      "learning_rate": 9.262630536456153e-06,
+      "loss": 0.8143,
+      "step": 520
+    },
+    {
+      "epoch": 0.2040619886418327,
+      "grad_norm": 0.06338963657617569,
+      "learning_rate": 9.229706346044749e-06,
+      "loss": 0.7735,
+      "step": 530
+    },
+    {
+      "epoch": 0.207912214842622,
+      "grad_norm": 0.06392891705036163,
+      "learning_rate": 9.196124273642422e-06,
+      "loss": 0.7949,
+      "step": 540
+    },
+    {
+      "epoch": 0.2117624410434113,
+      "grad_norm": 0.05758577585220337,
+      "learning_rate": 9.161889542552842e-06,
+      "loss": 0.8012,
+      "step": 550
+    },
+    {
+      "epoch": 0.21561266724420058,
+      "grad_norm": 0.06760473549365997,
+      "learning_rate": 9.127007477593193e-06,
+      "loss": 0.8019,
+      "step": 560
+    },
+    {
+      "epoch": 0.2194628934449899,
+      "grad_norm": 0.06395800411701202,
+      "learning_rate": 9.091483504265974e-06,
+      "loss": 0.8179,
+      "step": 570
+    },
+    {
+      "epoch": 0.22331311964577918,
+      "grad_norm": 0.06005476042628288,
+      "learning_rate": 9.055323147915116e-06,
+      "loss": 0.7863,
+      "step": 580
+    },
+    {
+      "epoch": 0.2271633458465685,
+      "grad_norm": 0.0633968710899353,
+      "learning_rate": 9.018532032866585e-06,
+      "loss": 0.8132,
+      "step": 590
+    },
+    {
+      "epoch": 0.23101357204735778,
+      "grad_norm": 0.0652826577425003,
+      "learning_rate": 8.981115881553579e-06,
+      "loss": 0.8048,
+      "step": 600
+    },
+    {
+      "epoch": 0.23486379824814707,
+      "grad_norm": 0.06125689297914505,
+      "learning_rate": 8.943080513626476e-06,
+      "loss": 0.8194,
+      "step": 610
+    },
+    {
+      "epoch": 0.23871402444893638,
+      "grad_norm": 0.06483124941587448,
+      "learning_rate": 8.904431845047641e-06,
+      "loss": 0.814,
+      "step": 620
+    },
+    {
+      "epoch": 0.24256425064972567,
+      "grad_norm": 0.07130519300699234,
+      "learning_rate": 8.86517588717128e-06,
+      "loss": 0.8107,
+      "step": 630
+    },
+    {
+      "epoch": 0.24641447685051496,
+      "grad_norm": 0.06208042800426483,
+      "learning_rate": 8.82531874580844e-06,
+      "loss": 0.8045,
+      "step": 640
+    },
+    {
+      "epoch": 0.25026470305130427,
+      "grad_norm": 0.0533473864197731,
+      "learning_rate": 8.784866620277314e-06,
+      "loss": 0.7709,
+      "step": 650
+    },
+    {
+      "epoch": 0.25411492925209356,
+      "grad_norm": 0.07049230486154556,
+      "learning_rate": 8.743825802439018e-06,
+      "loss": 0.8004,
+      "step": 660
+    },
+    {
+      "epoch": 0.25796515545288284,
+      "grad_norm": 0.06388885527849197,
+      "learning_rate": 8.702202675718956e-06,
+      "loss": 0.8025,
+      "step": 670
+    },
+    {
+      "epoch": 0.26181538165367213,
+      "grad_norm": 0.05791584402322769,
+      "learning_rate": 8.660003714113959e-06,
+      "loss": 0.7766,
+      "step": 680
+    },
+    {
+      "epoch": 0.26566560785446147,
+      "grad_norm": 0.05856083333492279,
+      "learning_rate": 8.617235481185323e-06,
+      "loss": 0.8112,
+      "step": 690
+    },
+    {
+      "epoch": 0.26951583405525076,
+      "grad_norm": 0.058459747582674026,
+      "learning_rate": 8.573904629037931e-06,
+      "loss": 0.7998,
+      "step": 700
+    },
+    {
+      "epoch": 0.27336606025604004,
+      "grad_norm": 0.058270908892154694,
+      "learning_rate": 8.530017897285585e-06,
+      "loss": 0.7923,
+      "step": 710
+    },
+    {
+      "epoch": 0.27721628645682933,
+      "grad_norm": 0.060456354171037674,
+      "learning_rate": 8.485582112002746e-06,
+      "loss": 0.7903,
+      "step": 720
+    },
+    {
+      "epoch": 0.2810665126576186,
+      "grad_norm": 0.06124652922153473,
+      "learning_rate": 8.440604184662811e-06,
+      "loss": 0.8009,
+      "step": 730
+    },
+    {
+      "epoch": 0.28491673885840796,
+      "grad_norm": 0.0653405487537384,
+      "learning_rate": 8.395091111063118e-06,
+      "loss": 0.7824,
+      "step": 740
+    },
+    {
+      "epoch": 0.28876696505919724,
+      "grad_norm": 0.05968422815203667,
+      "learning_rate": 8.349049970236822e-06,
+      "loss": 0.7869,
+      "step": 750
+    },
+    {
+      "epoch": 0.29261719125998653,
+      "grad_norm": 0.06110106036067009,
+      "learning_rate": 8.302487923351842e-06,
+      "loss": 0.7792,
+      "step": 760
+    },
+    {
+      "epoch": 0.2964674174607758,
+      "grad_norm": 0.0714542418718338,
+      "learning_rate": 8.255412212597021e-06,
+      "loss": 0.7923,
+      "step": 770
+    },
+    {
+      "epoch": 0.3003176436615651,
+      "grad_norm": 0.06303945928812027,
+      "learning_rate": 8.207830160055688e-06,
+      "loss": 0.7848,
+      "step": 780
+    },
+    {
+      "epoch": 0.3041678698623544,
+      "grad_norm": 0.06270917505025864,
+      "learning_rate": 8.159749166566786e-06,
+      "loss": 0.7832,
+      "step": 790
+    },
+    {
+      "epoch": 0.30801809606314373,
+      "grad_norm": 0.06209821254014969,
+      "learning_rate": 8.11117671057377e-06,
+      "loss": 0.8016,
+      "step": 800
+    },
+    {
+      "epoch": 0.311868322263933,
+      "grad_norm": 0.061749838292598724,
+      "learning_rate": 8.062120346961404e-06,
+      "loss": 0.7859,
+      "step": 810
+    },
+    {
+      "epoch": 0.3157185484647223,
+      "grad_norm": 0.05872559919953346,
+      "learning_rate": 8.012587705880705e-06,
+      "loss": 0.7956,
+      "step": 820
+    },
+    {
+      "epoch": 0.3195687746655116,
+      "grad_norm": 0.05891706794500351,
+      "learning_rate": 7.962586491562146e-06,
+      "loss": 0.7818,
+      "step": 830
+    },
+    {
+      "epoch": 0.3234190008663009,
+      "grad_norm": 0.06273745745420456,
+      "learning_rate": 7.912124481117357e-06,
+      "loss": 0.769,
+      "step": 840
+    },
+    {
+      "epoch": 0.3272692270670902,
+      "grad_norm": 0.06950563192367554,
+      "learning_rate": 7.861209523329494e-06,
+      "loss": 0.7888,
+      "step": 850
+    },
+    {
+      "epoch": 0.3311194532678795,
+      "grad_norm": 0.061616647988557816,
+      "learning_rate": 7.809849537432432e-06,
+      "loss": 0.7756,
+      "step": 860
+    },
+    {
+      "epoch": 0.3349696794686688,
+      "grad_norm": 0.056318264454603195,
+      "learning_rate": 7.758052511879035e-06,
+      "loss": 0.7774,
+      "step": 870
+    },
+    {
+      "epoch": 0.3388199056694581,
+      "grad_norm": 0.056205667555332184,
+      "learning_rate": 7.705826503098642e-06,
+      "loss": 0.7812,
+      "step": 880
+    },
+    {
+      "epoch": 0.34267013187024736,
+      "grad_norm": 0.06327362358570099,
+      "learning_rate": 7.653179634243979e-06,
+      "loss": 0.7694,
+      "step": 890
+    },
+    {
+      "epoch": 0.34652035807103665,
+      "grad_norm": 0.05637519434094429,
+      "learning_rate": 7.600120093927696e-06,
+      "loss": 0.7903,
+      "step": 900
+    },
+    {
+      "epoch": 0.350370584271826,
+      "grad_norm": 0.06297080963850021,
+      "learning_rate": 7.546656134948729e-06,
+      "loss": 0.7725,
+      "step": 910
+    },
+    {
+      "epoch": 0.3542208104726153,
+      "grad_norm": 0.06447892636060715,
+      "learning_rate": 7.492796073008661e-06,
+      "loss": 0.7662,
+      "step": 920
+    },
+    {
+      "epoch": 0.35807103667340456,
+      "grad_norm": 0.0703514888882637,
+      "learning_rate": 7.438548285418326e-06,
+      "loss": 0.7667,
+      "step": 930
+    },
+    {
+      "epoch": 0.36192126287419385,
+      "grad_norm": 0.06380635499954224,
+      "learning_rate": 7.3839212097948e-06,
+      "loss": 0.7903,
+      "step": 940
+    },
+    {
+      "epoch": 0.36577148907498314,
+      "grad_norm": 0.05741297826170921,
+      "learning_rate": 7.328923342749039e-06,
+      "loss": 0.8002,
+      "step": 950
+    },
+    {
+      "epoch": 0.3696217152757725,
+      "grad_norm": 0.06448666751384735,
+      "learning_rate": 7.273563238564319e-06,
+      "loss": 0.7807,
+      "step": 960
+    },
+    {
+      "epoch": 0.37347194147656176,
+      "grad_norm": 0.05759265273809433,
+      "learning_rate": 7.217849507865724e-06,
+      "loss": 0.7717,
+      "step": 970
+    },
+    {
+      "epoch": 0.37732216767735105,
+      "grad_norm": 0.059042394161224365,
+      "learning_rate": 7.161790816280858e-06,
+      "loss": 0.7745,
+      "step": 980
+    },
+    {
+      "epoch": 0.38117239387814034,
+      "grad_norm": 0.06329599022865295,
+      "learning_rate": 7.1053958830920125e-06,
+      "loss": 0.7726,
+      "step": 990
+    },
+    {
+      "epoch": 0.3850226200789296,
+      "grad_norm": 0.06019292399287224,
+      "learning_rate": 7.0486734798799715e-06,
+      "loss": 0.7864,
+      "step": 1000
+    },
+    {
+      "epoch": 0.3888728462797189,
+      "grad_norm": 0.05998087301850319,
+      "learning_rate": 6.991632429159708e-06,
+      "loss": 0.7931,
+      "step": 1010
+    },
+    {
+      "epoch": 0.39272307248050825,
+      "grad_norm": 0.06240907683968544,
+      "learning_rate": 6.934281603008132e-06,
+      "loss": 0.7892,
+      "step": 1020
+    },
+    {
+      "epoch": 0.39657329868129754,
+      "grad_norm": 0.058498937636613846,
+      "learning_rate": 6.876629921684143e-06,
+      "loss": 0.7933,
+      "step": 1030
+    },
+    {
+      "epoch": 0.4004235248820868,
+      "grad_norm": 0.06288294494152069,
+      "learning_rate": 6.818686352241199e-06,
+      "loss": 0.794,
+      "step": 1040
+    },
+    {
+      "epoch": 0.4042737510828761,
+      "grad_norm": 0.05776715651154518,
+      "learning_rate": 6.760459907132573e-06,
+      "loss": 0.7918,
+      "step": 1050
+    },
+    {
+      "epoch": 0.4081239772836654,
+      "grad_norm": 0.0582711324095726,
+      "learning_rate": 6.701959642809589e-06,
+      "loss": 0.7824,
+      "step": 1060
+    },
+    {
+      "epoch": 0.41197420348445474,
+      "grad_norm": 0.05871543660759926,
+      "learning_rate": 6.643194658312987e-06,
+      "loss": 0.7663,
+      "step": 1070
+    },
+    {
+      "epoch": 0.415824429685244,
+      "grad_norm": 0.06075074151158333,
+      "learning_rate": 6.584174093857676e-06,
+      "loss": 0.8062,
+      "step": 1080
+    },
+    {
+      "epoch": 0.4196746558860333,
+      "grad_norm": 0.06957266479730606,
+      "learning_rate": 6.5249071294110735e-06,
+      "loss": 0.773,
+      "step": 1090
+    },
+    {
+      "epoch": 0.4235248820868226,
+      "grad_norm": 0.06562729924917221,
+      "learning_rate": 6.465402983265274e-06,
+      "loss": 0.8015,
+      "step": 1100
+    },
+    {
+      "epoch": 0.4273751082876119,
+      "grad_norm": 0.06113231182098389,
+      "learning_rate": 6.405670910603243e-06,
+      "loss": 0.7943,
+      "step": 1110
+    },
+    {
+      "epoch": 0.43122533448840117,
+      "grad_norm": 0.06872578710317612,
+      "learning_rate": 6.345720202059291e-06,
+      "loss": 0.79,
+      "step": 1120
+    },
+    {
+      "epoch": 0.4350755606891905,
+      "grad_norm": 0.06504639983177185,
+      "learning_rate": 6.285560182274008e-06,
+      "loss": 0.7885,
+      "step": 1130
+    },
+    {
+      "epoch": 0.4389257868899798,
+      "grad_norm": 0.05811650678515434,
+      "learning_rate": 6.225200208443936e-06,
+      "loss": 0.7781,
+      "step": 1140
+    },
+    {
+      "epoch": 0.4427760130907691,
+      "grad_norm": 0.08444683998823166,
+      "learning_rate": 6.16464966886616e-06,
+      "loss": 0.7837,
+      "step": 1150
+    },
+    {
+      "epoch": 0.44662623929155837,
+      "grad_norm": 0.06788496673107147,
+      "learning_rate": 6.103917981478065e-06,
+      "loss": 0.8011,
+      "step": 1160
+    },
+    {
+      "epoch": 0.45047646549234766,
+      "grad_norm": 0.05947490409016609,
+      "learning_rate": 6.043014592392489e-06,
+      "loss": 0.8096,
+      "step": 1170
+    },
+    {
+      "epoch": 0.454326691693137,
+      "grad_norm": 0.06792280822992325,
+      "learning_rate": 5.981948974428485e-06,
+      "loss": 0.779,
+      "step": 1180
+    },
+    {
+      "epoch": 0.4581769178939263,
+      "grad_norm": 0.06291098892688751,
+      "learning_rate": 5.920730625637934e-06,
+      "loss": 0.7632,
+      "step": 1190
+    },
+    {
+      "epoch": 0.46202714409471557,
+      "grad_norm": 0.06250324100255966,
+      "learning_rate": 5.8593690678282355e-06,
+      "loss": 0.7745,
+      "step": 1200
+    },
+    {
+      "epoch": 0.46587737029550486,
+      "grad_norm": 0.0664227083325386,
+      "learning_rate": 5.797873845081301e-06,
+      "loss": 0.7746,
+      "step": 1210
+    },
+    {
+      "epoch": 0.46972759649629414,
+      "grad_norm": 0.06644179672002792,
+      "learning_rate": 5.736254522269075e-06,
+      "loss": 0.7974,
+      "step": 1220
+    },
+    {
+      "epoch": 0.47357782269708343,
+      "grad_norm": 0.0604996420443058,
+      "learning_rate": 5.6745206835658386e-06,
+      "loss": 0.7867,
+      "step": 1230
+    },
+    {
+      "epoch": 0.47742804889787277,
+      "grad_norm": 0.05940469726920128,
+      "learning_rate": 5.612681930957495e-06,
+      "loss": 0.7706,
+      "step": 1240
+    },
+    {
+      "epoch": 0.48127827509866206,
+      "grad_norm": 0.06329769641160965,
+      "learning_rate": 5.550747882748095e-06,
+      "loss": 0.7709,
+      "step": 1250
+    },
+    {
+      "epoch": 0.48512850129945134,
+      "grad_norm": 0.06472501903772354,
+      "learning_rate": 5.4887281720638195e-06,
+      "loss": 0.7748,
+      "step": 1260
+    },
+    {
+      "epoch": 0.48897872750024063,
+      "grad_norm": 0.06639126688241959,
+      "learning_rate": 5.42663244535466e-06,
+      "loss": 0.7624,
+      "step": 1270
+    },
+    {
+      "epoch": 0.4928289537010299,
+      "grad_norm": 0.05679118633270264,
+      "learning_rate": 5.3644703608940205e-06,
+      "loss": 0.7696,
+      "step": 1280
+    },
+    {
+      "epoch": 0.49667917990181926,
+      "grad_norm": 0.05756328999996185,
+      "learning_rate": 5.302251587276491e-06,
+      "loss": 0.7959,
+      "step": 1290
+    },
+    {
+      "epoch": 0.5005294061026085,
+      "grad_norm": 0.060336485505104065,
+      "learning_rate": 5.2399858019140005e-06,
+      "loss": 0.7821,
+      "step": 1300
+    },
+    {
+      "epoch": 0.5043796323033978,
+      "grad_norm": 0.06364555656909943,
+      "learning_rate": 5.177682689530615e-06,
+      "loss": 0.791,
+      "step": 1310
+    },
+    {
+      "epoch": 0.5082298585041871,
+      "grad_norm": 0.06056945025920868,
+      "learning_rate": 5.11535194065619e-06,
+      "loss": 0.7795,
+      "step": 1320
+    },
+    {
+      "epoch": 0.5120800847049765,
+      "grad_norm": 0.07635007798671722,
+      "learning_rate": 5.053003250119113e-06,
+      "loss": 0.7795,
+      "step": 1330
+    },
+    {
+      "epoch": 0.5159303109057657,
+      "grad_norm": 0.06159932538866997,
+      "learning_rate": 4.9906463155383925e-06,
+      "loss": 0.7855,
+      "step": 1340
+    },
+    {
+      "epoch": 0.519780537106555,
+      "grad_norm": 0.061512723565101624,
+      "learning_rate": 4.9282908358153046e-06,
+      "loss": 0.7715,
+      "step": 1350
+    },
+    {
+      "epoch": 0.5236307633073443,
+      "grad_norm": 0.0662158653140068,
+      "learning_rate": 4.8659465096248345e-06,
+      "loss": 0.7844,
+      "step": 1360
+    },
+    {
+      "epoch": 0.5274809895081336,
+      "grad_norm": 0.06481441110372543,
+      "learning_rate": 4.803623033907169e-06,
+      "loss": 0.7877,
+      "step": 1370
+    },
+    {
+      "epoch": 0.5313312157089229,
+      "grad_norm": 0.055871278047561646,
+      "learning_rate": 4.741330102359441e-06,
+      "loss": 0.7981,
+      "step": 1380
+    },
+    {
+      "epoch": 0.5351814419097122,
+      "grad_norm": 0.06593842804431915,
+      "learning_rate": 4.679077403927996e-06,
+      "loss": 0.7709,
+      "step": 1390
+    },
+    {
+      "epoch": 0.5390316681105015,
+      "grad_norm": 0.0638933852314949,
+      "learning_rate": 4.616874621301373e-06,
+      "loss": 0.7726,
+      "step": 1400
+    },
+    {
+      "epoch": 0.5428818943112907,
+      "grad_norm": 0.06753615289926529,
+      "learning_rate": 4.554731429404293e-06,
+      "loss": 0.7972,
+      "step": 1410
+    },
+    {
+      "epoch": 0.5467321205120801,
+      "grad_norm": 0.06624545902013779,
+      "learning_rate": 4.492657493892822e-06,
+      "loss": 0.7777,
+      "step": 1420
+    },
+    {
+      "epoch": 0.5505823467128694,
+      "grad_norm": 0.06724836677312851,
+      "learning_rate": 4.430662469650999e-06,
+      "loss": 0.7831,
+      "step": 1430
+    },
+    {
+      "epoch": 0.5544325729136587,
+      "grad_norm": 0.06063222512602806,
+      "learning_rate": 4.368755999289118e-06,
+      "loss": 0.7968,
+      "step": 1440
+    },
+    {
+      "epoch": 0.558282799114448,
+      "grad_norm": 0.06383723765611649,
+      "learning_rate": 4.306947711643949e-06,
+      "loss": 0.7768,
+      "step": 1450
+    },
+    {
+      "epoch": 0.5621330253152372,
+      "grad_norm": 0.0642712414264679,
+      "learning_rate": 4.245247220281068e-06,
+      "loss": 0.7796,
+      "step": 1460
+    },
+    {
+      "epoch": 0.5659832515160266,
+      "grad_norm": 0.06087171286344528,
+      "learning_rate": 4.183664121999587e-06,
+      "loss": 0.7741,
+      "step": 1470
+    },
+    {
+      "epoch": 0.5698334777168159,
+      "grad_norm": 0.06361310184001923,
+      "learning_rate": 4.122207995339474e-06,
+      "loss": 0.7701,
+      "step": 1480
+    },
+    {
+      "epoch": 0.5736837039176051,
+      "grad_norm": 0.06855567544698715,
+      "learning_rate": 4.060888399091734e-06,
+      "loss": 0.7785,
+      "step": 1490
+    },
+    {
+      "epoch": 0.5775339301183945,
+      "grad_norm": 0.05951743200421333,
+      "learning_rate": 3.999714870811639e-06,
+      "loss": 0.7918,
+      "step": 1500
+    },
+    {
+      "epoch": 0.5813841563191837,
+      "grad_norm": 0.06631240248680115,
+      "learning_rate": 3.938696925335275e-06,
+      "loss": 0.7908,
+      "step": 1510
+    },
+    {
+      "epoch": 0.5852343825199731,
+      "grad_norm": 0.06370088458061218,
+      "learning_rate": 3.87784405329962e-06,
+      "loss": 0.7877,
+      "step": 1520
+    },
+    {
+      "epoch": 0.5890846087207623,
+      "grad_norm": 0.06418631970882416,
+      "learning_rate": 3.817165719666382e-06,
+      "loss": 0.796,
+      "step": 1530
+    },
+    {
+      "epoch": 0.5929348349215516,
+      "grad_norm": 0.05913081392645836,
+      "learning_rate": 3.756671362249841e-06,
+      "loss": 0.79,
+      "step": 1540
+    },
+    {
+      "epoch": 0.596785061122341,
+      "grad_norm": 0.06324242800474167,
+      "learning_rate": 3.6963703902488907e-06,
+      "loss": 0.7963,
+      "step": 1550
+    },
+    {
+      "epoch": 0.6006352873231302,
+      "grad_norm": 0.06874449551105499,
+      "learning_rate": 3.6362721827835655e-06,
+      "loss": 0.7995,
+      "step": 1560
+    },
+    {
+      "epoch": 0.6044855135239195,
+      "grad_norm": 0.06754579395055771,
+      "learning_rate": 3.576386087436212e-06,
+      "loss": 0.7819,
+      "step": 1570
+    },
+    {
+      "epoch": 0.6083357397247088,
+      "grad_norm": 0.06520285457372665,
+      "learning_rate": 3.5167214187975913e-06,
+      "loss": 0.7858,
+      "step": 1580
+    },
+    {
+      "epoch": 0.6121859659254981,
+      "grad_norm": 0.05465632304549217,
+      "learning_rate": 3.457287457018086e-06,
+      "loss": 0.7945,
+      "step": 1590
+    },
+    {
+      "epoch": 0.6160361921262875,
+      "grad_norm": 0.06167367473244667,
+      "learning_rate": 3.3980934463643003e-06,
+      "loss": 0.7789,
+      "step": 1600
+    },
+    {
+      "epoch": 0.6198864183270767,
+      "grad_norm": 0.05676029995083809,
+      "learning_rate": 3.3391485937812053e-06,
+      "loss": 0.7565,
+      "step": 1610
+    },
+    {
+      "epoch": 0.623736644527866,
+      "grad_norm": 0.059358615428209305,
+      "learning_rate": 3.280462067460115e-06,
+      "loss": 0.7556,
+      "step": 1620
+    },
+    {
+      "epoch": 0.6275868707286553,
+      "grad_norm": 0.05931363254785538,
+      "learning_rate": 3.222042995412669e-06,
+      "loss": 0.7887,
+      "step": 1630
+    },
+    {
+      "epoch": 0.6314370969294446,
+      "grad_norm": 0.052229367196559906,
+      "learning_rate": 3.163900464051089e-06,
+      "loss": 0.7742,
+      "step": 1640
+    },
+    {
+      "epoch": 0.635287323130234,
+      "grad_norm": 0.06011379882693291,
+      "learning_rate": 3.1060435167748826e-06,
+      "loss": 0.7839,
+      "step": 1650
+    },
+    {
+      "epoch": 0.6391375493310232,
+      "grad_norm": 0.05692354217171669,
+      "learning_rate": 3.048481152564253e-06,
+      "loss": 0.7643,
+      "step": 1660
+    },
+    {
+      "epoch": 0.6429877755318125,
+      "grad_norm": 0.06393871456384659,
+      "learning_rate": 2.9912223245804017e-06,
+      "loss": 0.7901,
+      "step": 1670
+    },
+    {
+      "epoch": 0.6468380017326018,
+      "grad_norm": 0.06268513947725296,
+      "learning_rate": 2.934275938772981e-06,
+      "loss": 0.7784,
+      "step": 1680
+    },
+    {
+      "epoch": 0.6506882279333911,
+      "grad_norm": 0.0566086545586586,
+      "learning_rate": 2.8776508524948676e-06,
+      "loss": 0.7718,
+      "step": 1690
+    },
+    {
+      "epoch": 0.6545384541341804,
+      "grad_norm": 0.05910525470972061,
+      "learning_rate": 2.821355873124503e-06,
+      "loss": 0.7725,
+      "step": 1700
+    },
+    {
+      "epoch": 0.6583886803349697,
+      "grad_norm": 0.06373787671327591,
+      "learning_rate": 2.7653997566960144e-06,
+      "loss": 0.7867,
+      "step": 1710
+    },
+    {
+      "epoch": 0.662238906535759,
+      "grad_norm": 0.06430958211421967,
+      "learning_rate": 2.709791206537309e-06,
+      "loss": 0.7749,
+      "step": 1720
+    },
+    {
+      "epoch": 0.6660891327365482,
+      "grad_norm": 0.05788276717066765,
+      "learning_rate": 2.6545388719163747e-06,
+      "loss": 0.7815,
+      "step": 1730
+    },
+    {
+      "epoch": 0.6699393589373376,
+      "grad_norm": 0.06402074545621872,
+      "learning_rate": 2.599651346695979e-06,
+      "loss": 0.7822,
+      "step": 1740
+    },
+    {
+      "epoch": 0.6737895851381268,
+      "grad_norm": 0.05772216245532036,
+      "learning_rate": 2.545137167997007e-06,
+      "loss": 0.7855,
+      "step": 1750
+    },
+    {
+      "epoch": 0.6776398113389162,
+      "grad_norm": 0.05752065032720566,
+      "learning_rate": 2.4910048148705972e-06,
+      "loss": 0.7744,
+      "step": 1760
+    },
+    {
+      "epoch": 0.6814900375397055,
+      "grad_norm": 0.06220688298344612,
+      "learning_rate": 2.4372627069793327e-06,
+      "loss": 0.8042,
+      "step": 1770
+    },
+    {
+      "epoch": 0.6853402637404947,
+      "grad_norm": 0.062244147062301636,
+      "learning_rate": 2.3839192032876486e-06,
+      "loss": 0.792,
+      "step": 1780
+    },
+    {
+      "epoch": 0.6891904899412841,
+      "grad_norm": 0.05976283922791481,
+      "learning_rate": 2.3309826007617047e-06,
+      "loss": 0.8063,
+      "step": 1790
+    },
+    {
+      "epoch": 0.6930407161420733,
+      "grad_norm": 0.06013132631778717,
+      "learning_rate": 2.278461133078881e-06,
+      "loss": 0.7661,
+      "step": 1800
+    },
+    {
+      "epoch": 0.6968909423428626,
+      "grad_norm": 0.06046042591333389,
+      "learning_rate": 2.2263629693471258e-06,
+      "loss": 0.7743,
+      "step": 1810
+    },
+    {
+      "epoch": 0.700741168543652,
+      "grad_norm": 0.06628702580928802,
+      "learning_rate": 2.1746962128343414e-06,
+      "loss": 0.7927,
+      "step": 1820
+    },
+    {
+      "epoch": 0.7045913947444412,
+      "grad_norm": 0.05554218217730522,
+      "learning_rate": 2.1234688997080183e-06,
+      "loss": 0.7852,
+      "step": 1830
+    },
+    {
+      "epoch": 0.7084416209452306,
+      "grad_norm": 0.06405815482139587,
+      "learning_rate": 2.0726889977853053e-06,
+      "loss": 0.7808,
+      "step": 1840
+    },
+    {
+      "epoch": 0.7122918471460198,
+      "grad_norm": 0.06627866625785828,
+      "learning_rate": 2.022364405293703e-06,
+      "loss": 0.7841,
+      "step": 1850
+    },
+    {
+      "epoch": 0.7161420733468091,
+      "grad_norm": 0.06092619523406029,
+      "learning_rate": 1.9725029496425814e-06,
+      "loss": 0.7909,
+      "step": 1860
+    },
+    {
+      "epoch": 0.7199922995475985,
+      "grad_norm": 0.05669494718313217,
+      "learning_rate": 1.923112386205727e-06,
+      "loss": 0.783,
+      "step": 1870
+    },
+    {
+      "epoch": 0.7238425257483877,
+      "grad_norm": 0.06416063010692596,
+      "learning_rate": 1.874200397115079e-06,
+      "loss": 0.7886,
+      "step": 1880
+    },
+    {
+      "epoch": 0.727692751949177,
+      "grad_norm": 0.06330247968435287,
+      "learning_rate": 1.825774590065857e-06,
+      "loss": 0.7619,
+      "step": 1890
+    },
+    {
+      "epoch": 0.7315429781499663,
+      "grad_norm": 0.06259039789438248,
+      "learning_rate": 1.7778424971332813e-06,
+      "loss": 0.7892,
+      "step": 1900
+    },
+    {
+      "epoch": 0.7353932043507556,
+      "grad_norm": 0.057196132838726044,
+      "learning_rate": 1.7304115736010407e-06,
+      "loss": 0.7851,
+      "step": 1910
+    },
+    {
+      "epoch": 0.739243430551545,
+      "grad_norm": 0.06251382827758789,
+      "learning_rate": 1.6834891968017108e-06,
+      "loss": 0.7683,
+      "step": 1920
+    },
+    {
+      "epoch": 0.7430936567523342,
+      "grad_norm": 0.0644979476928711,
+      "learning_rate": 1.6370826649692884e-06,
+      "loss": 0.7944,
+      "step": 1930
+    },
+    {
+      "epoch": 0.7469438829531235,
+      "grad_norm": 0.06054569408297539,
+      "learning_rate": 1.5911991961040463e-06,
+      "loss": 0.7896,
+      "step": 1940
+    },
+    {
+      "epoch": 0.7507941091539128,
+      "grad_norm": 0.058386821299791336,
+      "learning_rate": 1.545845926849851e-06,
+      "loss": 0.7845,
+      "step": 1950
+    },
+    {
+      "epoch": 0.7546443353547021,
+      "grad_norm": 0.062304794788360596,
+      "learning_rate": 1.5010299113841397e-06,
+      "loss": 0.8081,
+      "step": 1960
+    },
+    {
+      "epoch": 0.7584945615554913,
+      "grad_norm": 0.058821115642786026,
+      "learning_rate": 1.4567581203207238e-06,
+      "loss": 0.7853,
+      "step": 1970
+    },
+    {
+      "epoch": 0.7623447877562807,
+      "grad_norm": 0.06243539974093437,
+      "learning_rate": 1.413037439625593e-06,
+      "loss": 0.7776,
+      "step": 1980
+    },
+    {
+      "epoch": 0.76619501395707,
+      "grad_norm": 0.06881844997406006,
+      "learning_rate": 1.3698746695458847e-06,
+      "loss": 0.7923,
+      "step": 1990
+    },
+    {
+      "epoch": 0.7700452401578592,
+      "grad_norm": 0.056873828172683716,
+      "learning_rate": 1.3272765235521745e-06,
+      "loss": 0.7761,
+      "step": 2000
+    },
+    {
+      "epoch": 0.7738954663586486,
+      "grad_norm": 0.06489020586013794,
+      "learning_rate": 1.2852496272942877e-06,
+      "loss": 0.7798,
+      "step": 2010
+    },
+    {
+      "epoch": 0.7777456925594378,
+      "grad_norm": 0.05470595136284828,
+      "learning_rate": 1.2438005175707396e-06,
+      "loss": 0.7722,
+      "step": 2020
+    },
+    {
+      "epoch": 0.7815959187602272,
+      "grad_norm": 0.06360286474227905,
+      "learning_rate": 1.2029356413120297e-06,
+      "loss": 0.7663,
+      "step": 2030
+    },
+    {
+      "epoch": 0.7854461449610165,
+      "grad_norm": 0.0612981840968132,
+      "learning_rate": 1.1626613545778786e-06,
+      "loss": 0.782,
+      "step": 2040
+    },
+    {
+      "epoch": 0.7892963711618057,
+      "grad_norm": 0.0599975623190403,
+      "learning_rate": 1.1229839215686294e-06,
+      "loss": 0.7816,
+      "step": 2050
+    },
+    {
+      "epoch": 0.7931465973625951,
+      "grad_norm": 0.05607706680893898,
+      "learning_rate": 1.0839095136509188e-06,
+      "loss": 0.7663,
+      "step": 2060
+    },
+    {
+      "epoch": 0.7969968235633843,
+      "grad_norm": 0.06426962465047836,
+      "learning_rate": 1.045444208397791e-06,
+      "loss": 0.7889,
+      "step": 2070
+    },
+    {
+      "epoch": 0.8008470497641736,
+      "grad_norm": 0.05778040364384651,
+      "learning_rate": 1.007593988643402e-06,
+      "loss": 0.761,
+      "step": 2080
+    },
+    {
+      "epoch": 0.804697275964963,
+      "grad_norm": 0.05679520219564438,
+      "learning_rate": 9.703647415524613e-07,
+      "loss": 0.7843,
+      "step": 2090
+    },
+    {
+      "epoch": 0.8085475021657522,
+      "grad_norm": 0.06228042393922806,
+      "learning_rate": 9.337622577045485e-07,
+      "loss": 0.7854,
+      "step": 2100
+    },
+    {
+      "epoch": 0.8123977283665416,
+      "grad_norm": 0.06188025325536728,
+      "learning_rate": 8.977922301934588e-07,
+      "loss": 0.7819,
+      "step": 2110
+    },
+    {
+      "epoch": 0.8162479545673308,
+      "grad_norm": 0.059744417667388916,
+      "learning_rate": 8.624602537417009e-07,
+      "loss": 0.7814,
+      "step": 2120
+    },
+    {
+      "epoch": 0.8200981807681201,
+      "grad_norm": 0.06239009276032448,
+      "learning_rate": 8.277718238303123e-07,
+      "loss": 0.7784,
+      "step": 2130
+    },
+    {
+      "epoch": 0.8239484069689095,
+      "grad_norm": 0.05940278619527817,
+      "learning_rate": 7.937323358440935e-07,
+      "loss": 0.7819,
+      "step": 2140
+    },
+    {
+      "epoch": 0.8277986331696987,
+      "grad_norm": 0.07342144846916199,
+      "learning_rate": 7.603470842324223e-07,
+      "loss": 0.7775,
+      "step": 2150
+    },
+    {
+      "epoch": 0.831648859370488,
+      "grad_norm": 0.059887390583753586,
+      "learning_rate": 7.276212616857575e-07,
+      "loss": 0.7926,
+      "step": 2160
+    },
+    {
+      "epoch": 0.8354990855712773,
+      "grad_norm": 0.06092890724539757,
+      "learning_rate": 6.955599583279837e-07,
+      "loss": 0.7942,
+      "step": 2170
+    },
+    {
+      "epoch": 0.8393493117720666,
+      "grad_norm": 0.056721363216638565,
+      "learning_rate": 6.641681609246981e-07,
+      "loss": 0.7766,
+      "step": 2180
+    },
+    {
+      "epoch": 0.8431995379728559,
+      "grad_norm": 0.060754671692848206,
+      "learning_rate": 6.334507521075711e-07,
+      "loss": 0.7845,
+      "step": 2190
+    },
+    {
+      "epoch": 0.8470497641736452,
+      "grad_norm": 0.06513000279664993,
+      "learning_rate": 6.034125096149179e-07,
+      "loss": 0.7759,
+      "step": 2200
+    },
+    {
+      "epoch": 0.8508999903744345,
+      "grad_norm": 0.06753235310316086,
+      "learning_rate": 5.740581055485722e-07,
+      "loss": 0.7581,
+      "step": 2210
+    },
+    {
+      "epoch": 0.8547502165752238,
+      "grad_norm": 0.052786633372306824,
+      "learning_rate": 5.453921056471961e-07,
+      "loss": 0.7917,
+      "step": 2220
+    },
+    {
+      "epoch": 0.8586004427760131,
+      "grad_norm": 0.06240396574139595,
+      "learning_rate": 5.174189685761288e-07,
+      "loss": 0.7694,
+      "step": 2230
+    },
+    {
+      "epoch": 0.8624506689768023,
+      "grad_norm": 0.0601973757147789,
+      "learning_rate": 4.901430452338979e-07,
+      "loss": 0.7921,
+      "step": 2240
+    },
+    {
+      "epoch": 0.8663008951775917,
+      "grad_norm": 0.059187259525060654,
+      "learning_rate": 4.6356857807548286e-07,
+      "loss": 0.779,
+      "step": 2250
+    },
+    {
+      "epoch": 0.870151121378381,
+      "grad_norm": 0.06149625405669212,
+      "learning_rate": 4.376997004524541e-07,
+      "loss": 0.7715,
+      "step": 2260
+    },
+    {
+      "epoch": 0.8740013475791703,
+      "grad_norm": 0.06919949501752853,
+      "learning_rate": 4.12540435970073e-07,
+      "loss": 0.7809,
+      "step": 2270
+    },
+    {
+      "epoch": 0.8778515737799596,
+      "grad_norm": 0.05556255206465721,
+      "learning_rate": 3.880946978614703e-07,
+      "loss": 0.7818,
+      "step": 2280
+    },
+    {
+      "epoch": 0.8817017999807488,
+      "grad_norm": 0.05759495124220848,
+      "learning_rate": 3.643662883789878e-07,
+      "loss": 0.7957,
+      "step": 2290
+    },
+    {
+      "epoch": 0.8855520261815382,
+      "grad_norm": 0.055562037974596024,
+      "learning_rate": 3.41358898202781e-07,
+      "loss": 0.7714,
+      "step": 2300
+    },
+    {
+      "epoch": 0.8894022523823275,
+      "grad_norm": 0.05814225599169731,
+      "learning_rate": 3.190761058667741e-07,
+      "loss": 0.7808,
+      "step": 2310
+    },
+    {
+      "epoch": 0.8932524785831167,
+      "grad_norm": 0.05729026719927788,
+      "learning_rate": 2.9752137720206463e-07,
+      "loss": 0.7854,
+      "step": 2320
+    },
+    {
+      "epoch": 0.8971027047839061,
+      "grad_norm": 0.06092577427625656,
+      "learning_rate": 2.7669806479785067e-07,
+      "loss": 0.7933,
+      "step": 2330
+    },
+    {
+      "epoch": 0.9009529309846953,
+      "grad_norm": 0.05808475986123085,
+      "learning_rate": 2.566094074799763e-07,
+      "loss": 0.7796,
+      "step": 2340
+    },
+    {
+      "epoch": 0.9048031571854847,
+      "grad_norm": 0.05841906741261482,
+      "learning_rate": 2.3725852980716557e-07,
+      "loss": 0.7638,
+      "step": 2350
+    },
+    {
+      "epoch": 0.908653383386274,
+      "grad_norm": 0.06437428295612335,
+      "learning_rate": 2.1864844158503983e-07,
+      "loss": 0.7853,
+      "step": 2360
+    },
+    {
+      "epoch": 0.9125036095870632,
+      "grad_norm": 0.06203237175941467,
+      "learning_rate": 2.007820373979702e-07,
+      "loss": 0.765,
+      "step": 2370
+    },
+    {
+      "epoch": 0.9163538357878526,
+      "grad_norm": 0.061947330832481384,
+      "learning_rate": 1.8366209615886532e-07,
+      "loss": 0.7786,
+      "step": 2380
+    },
+    {
+      "epoch": 0.9202040619886418,
+      "grad_norm": 0.05867836996912956,
+      "learning_rate": 1.6729128067693613e-07,
+      "loss": 0.796,
+      "step": 2390
+    },
+    {
+      "epoch": 0.9240542881894311,
+      "grad_norm": 0.0602993406355381,
+      "learning_rate": 1.5167213724353426e-07,
+      "loss": 0.7796,
+      "step": 2400
+    },
+    {
+      "epoch": 0.9279045143902205,
+      "grad_norm": 0.062359053641557693,
+      "learning_rate": 1.3680709523610058e-07,
+      "loss": 0.7677,
+      "step": 2410
+    },
+    {
+      "epoch": 0.9317547405910097,
+      "grad_norm": 0.05947231873869896,
+      "learning_rate": 1.2269846674030418e-07,
+      "loss": 0.8004,
+      "step": 2420
+    },
+    {
+      "epoch": 0.935604966791799,
+      "grad_norm": 0.060753725469112396,
+      "learning_rate": 1.0934844619042595e-07,
+      "loss": 0.7887,
+      "step": 2430
+    },
+    {
+      "epoch": 0.9394551929925883,
+      "grad_norm": 0.06487417221069336,
+      "learning_rate": 9.675911002803961e-08,
+      "loss": 0.7765,
+      "step": 2440
+    },
+    {
+      "epoch": 0.9433054191933776,
+      "grad_norm": 0.061570022255182266,
+      "learning_rate": 8.493241637904381e-08,
+      "loss": 0.7726,
+      "step": 2450
+    },
+    {
+      "epoch": 0.9471556453941669,
+      "grad_norm": 0.060091178864240646,
+      "learning_rate": 7.387020474909689e-08,
+      "loss": 0.7975,
+      "step": 2460
+    },
+    {
+      "epoch": 0.9510058715949562,
+      "grad_norm": 0.05691954120993614,
+      "learning_rate": 6.357419573750856e-08,
+      "loss": 0.7759,
+      "step": 2470
+    },
+    {
+      "epoch": 0.9548560977957455,
+      "grad_norm": 0.05996343493461609,
+      "learning_rate": 5.4045990769613365e-08,
+      "loss": 0.7822,
+      "step": 2480
+    },
+    {
+      "epoch": 0.9587063239965348,
+      "grad_norm": 0.05826522037386894,
+      "learning_rate": 4.5287071847694434e-08,
+      "loss": 0.7803,
+      "step": 2490
+    },
+    {
+      "epoch": 0.9625565501973241,
+      "grad_norm": 0.06268733739852905,
+      "learning_rate": 3.7298801320467817e-08,
+      "loss": 0.7702,
+      "step": 2500
+    },
+    {
+      "epoch": 0.9664067763981133,
+      "grad_norm": 0.05827909708023071,
+      "learning_rate": 3.0082421671192576e-08,
+      "loss": 0.7806,
+      "step": 2510
+    },
+    {
+      "epoch": 0.9702570025989027,
+      "grad_norm": 0.059483546763658524,
+      "learning_rate": 2.3639055324410355e-08,
+      "loss": 0.7756,
+      "step": 2520
+    },
+    {
+      "epoch": 0.974107228799692,
+      "grad_norm": 0.0688377246260643,
+      "learning_rate": 1.7969704471369477e-08,
+      "loss": 0.7715,
+      "step": 2530
+    },
+    {
+      "epoch": 0.9779574550004813,
+      "grad_norm": 0.05891785770654678,
+      "learning_rate": 1.3075250914142413e-08,
+      "loss": 0.7785,
+      "step": 2540
+    },
+    {
+      "epoch": 0.9818076812012706,
+      "grad_norm": 0.057990189641714096,
+      "learning_rate": 8.956455928476049e-09,
+      "loss": 0.7904,
+      "step": 2550
+    },
+    {
+      "epoch": 0.9856579074020598,
+      "grad_norm": 0.054882582277059555,
+      "learning_rate": 5.6139601453786275e-09,
+      "loss": 0.7682,
+      "step": 2560
+    },
+    {
+      "epoch": 0.9895081336028492,
+      "grad_norm": 0.060419563204050064,
+      "learning_rate": 3.0482834514805602e-09,
+      "loss": 0.7676,
+      "step": 2570
+    },
+    {
+      "epoch": 0.9933583598036385,
+      "grad_norm": 0.058154281228780746,
+      "learning_rate": 1.2598249081713366e-09,
+      "loss": 0.7573,
+      "step": 2580
+    },
+    {
+      "epoch": 0.9972085860044277,
+      "grad_norm": 0.05702221766114235,
+      "learning_rate": 2.488626895302826e-10,
+      "loss": 0.7856,
+      "step": 2590
+    },
+    {
+      "epoch": 0.9999037443449803,
+      "step": 2597,
+      "total_flos": 2.877231303135068e+18,
+      "train_loss": 0.8017009085859754,
+      "train_runtime": 17598.5398,
+      "train_samples_per_second": 37.782,
+      "train_steps_per_second": 0.148
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 2597,
+  "num_input_tokens_seen": 246462940,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.877231303135068e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/training_args.bin ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:56f9783487ab921346a513c3910ba95e5c5ef0dc5919a885e04d0b3b745a8db1
+size 6264

continuous/qwen25/pixio_vith16/finetune.bad_hf_native_20260818/vocab.json ADDED Viewed

The diff for this file is too large to render. See raw diff

continuous/qwen25/pixio_vith16/finetune/added_tokens.json ADDED Viewed

	@@ -0,0 +1,24 @@

+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

continuous/qwen25/pixio_vith16/finetune/chat_template.jinja ADDED Viewed

	@@ -0,0 +1,54 @@

+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- messages[0]['content'] }}
+    {%- else %}
+        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
+    {%- endif %}
+    {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
+    {%- else %}
+        {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role }}
+        {%- if message.content %}
+            {{- '\n' + message.content }}
+        {%- endif %}
+        {%- for tool_call in message.tool_calls %}
+            {%- if tool_call.function is defined %}
+                {%- set tool_call = tool_call.function %}
+            {%- endif %}
+            {{- '\n<tool_call>\n{"name": "' }}
+            {{- tool_call.name }}
+            {{- '", "arguments": ' }}
+            {{- tool_call.arguments | tojson }}
+            {{- '}\n</tool_call>' }}
+        {%- endfor %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+{%- endif %}

continuous/qwen25/pixio_vith16/finetune/config.json ADDED Viewed

	@@ -0,0 +1,87 @@

+{
+  "add_faster_video": false,
+  "add_time_instruction": false,
+  "architectures": [
+    "LlavaQwenForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "dtype": "bfloat16",
+  "eos_token_id": 151645,
+  "faster_token_stride": 10,
+  "force_sample": false,
+  "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,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "mm_hidden_size": 1280,
+  "mm_newline_position": "grid",
+  "mm_patch_merge_type": "flat",
+  "mm_projector_lr": 0.0001,
+  "mm_projector_type": "mlp2x_gelu",
+  "mm_resampler_type": null,
+  "mm_spatial_pool_mode": "bilinear",
+  "mm_spatial_pool_stride": null,
+  "mm_tunable_parts": "mm_mlp_adapter,mm_language_model",
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -1,
+  "mm_vision_tower": "vtb_ssl:pixio",
+  "mm_vision_tower_lr": null,
+  "model_type": "qwen2",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 2,
+  "pad_token_id": 151643,
+  "pos_skipping_range": 4096,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "tokenizer_model_max_length": 8192,
+  "tokenizer_padding_side": "right",
+  "transformers_version": "4.57.6",
+  "use_cache": true,
+  "use_mm_proj": true,
+  "use_pos_skipping": false,
+  "use_sliding_window": false,
+  "vision_tower_pretrained": "/cache/ckpt/download/tokenizer/continuous/pixio/pixio-vith16/pixio_vith16.pth",
+  "vocab_size": 151936
+}

continuous/qwen25/pixio_vith16/finetune/generation_config.json ADDED Viewed

	@@ -0,0 +1,13 @@

+{
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.1,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "4.57.6"
+}

continuous/qwen25/pixio_vith16/finetune/merges.txt ADDED Viewed

The diff for this file is too large to render. See raw diff

continuous/qwen25/pixio_vith16/finetune/model.safetensors ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:599760f5d1d169fa2bf11dc925525f3d04d1d3a8144b5ec03ed3fda50fd46c78
+size 4833605896

continuous/qwen25/pixio_vith16/finetune/special_tokens_map.json ADDED Viewed

	@@ -0,0 +1,31 @@

+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

continuous/qwen25/pixio_vith16/finetune/tokenizer.json ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:9c5ae00e602b8860cbd784ba82a8aa14e8feecec692e7076590d014d7b7fdafa
+size 11421896

continuous/qwen25/pixio_vith16/finetune/tokenizer_config.json ADDED Viewed

	@@ -0,0 +1,208 @@

+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 8192,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

continuous/qwen25/pixio_vith16/finetune/trainer_state.json ADDED Viewed

	@@ -0,0 +1,1856 @@

+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9999037443449803,
+  "eval_steps": 500,
+  "global_step": 2597,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.003850226200789296,
+      "grad_norm": 5.219594955444336,
+      "learning_rate": 1.153846153846154e-06,
+      "loss": 2.0261,
+      "step": 10
+    },
+    {
+      "epoch": 0.007700452401578592,
+      "grad_norm": 3.4460830688476562,
+      "learning_rate": 2.435897435897436e-06,
+      "loss": 1.5673,
+      "step": 20
+    },
+    {
+      "epoch": 0.01155067860236789,
+      "grad_norm": 1.2486481666564941,
+      "learning_rate": 3.7179487179487184e-06,
+      "loss": 1.1625,
+      "step": 30
+    },
+    {
+      "epoch": 0.015400904803157185,
+      "grad_norm": 0.35704895853996277,
+      "learning_rate": 5e-06,
+      "loss": 0.9789,
+      "step": 40
+    },
+    {
+      "epoch": 0.019251131003946483,
+      "grad_norm": 0.16484658420085907,
+      "learning_rate": 6.282051282051282e-06,
+      "loss": 0.9353,
+      "step": 50
+    },
+    {
+      "epoch": 0.02310135720473578,
+      "grad_norm": 0.12496022880077362,
+      "learning_rate": 7.564102564102564e-06,
+      "loss": 0.8863,
+      "step": 60
+    },
+    {
+      "epoch": 0.026951583405525073,
+      "grad_norm": 0.09574976563453674,
+      "learning_rate": 8.846153846153847e-06,
+      "loss": 0.8588,
+      "step": 70
+    },
+    {
+      "epoch": 0.03080180960631437,
+      "grad_norm": 0.11452022194862366,
+      "learning_rate": 9.999996111488723e-06,
+      "loss": 0.8762,
+      "step": 80
+    },
+    {
+      "epoch": 0.034652035807103666,
+      "grad_norm": 0.09458762407302856,
+      "learning_rate": 9.999529497453782e-06,
+      "loss": 0.8529,
+      "step": 90
+    },
+    {
+      "epoch": 0.038502262007892966,
+      "grad_norm": 0.08667493611574173,
+      "learning_rate": 9.998285264324193e-06,
+      "loss": 0.8311,
+      "step": 100
+    },
+    {
+      "epoch": 0.04235248820868226,
+      "grad_norm": 0.09068763256072998,
+      "learning_rate": 9.996263605626052e-06,
+      "loss": 0.8312,
+      "step": 110
+    },
+    {
+      "epoch": 0.04620271440947156,
+      "grad_norm": 0.09635454416275024,
+      "learning_rate": 9.99346483580503e-06,
+      "loss": 0.8251,
+      "step": 120
+    },
+    {
+      "epoch": 0.05005294061026085,
+      "grad_norm": 0.07572688162326813,
+      "learning_rate": 9.989889390177467e-06,
+      "loss": 0.8129,
+      "step": 130
+    },
+    {
+      "epoch": 0.053903166811050146,
+      "grad_norm": 0.08203379809856415,
+      "learning_rate": 9.985537824862647e-06,
+      "loss": 0.8162,
+      "step": 140
+    },
+    {
+      "epoch": 0.057753393011839446,
+      "grad_norm": 0.09237272292375565,
+      "learning_rate": 9.980410816696317e-06,
+      "loss": 0.8187,
+      "step": 150
+    },
+    {
+      "epoch": 0.06160361921262874,
+      "grad_norm": 0.08760576695203781,
+      "learning_rate": 9.97450916312541e-06,
+      "loss": 0.8183,
+      "step": 160
+    },
+    {
+      "epoch": 0.06545384541341803,
+      "grad_norm": 0.08265096694231033,
+      "learning_rate": 9.967833782084004e-06,
+      "loss": 0.7976,
+      "step": 170
+    },
+    {
+      "epoch": 0.06930407161420733,
+      "grad_norm": 0.08389075100421906,
+      "learning_rate": 9.960385711850554e-06,
+      "loss": 0.8326,
+      "step": 180
+    },
+    {
+      "epoch": 0.07315429781499663,
+      "grad_norm": 0.0846567153930664,
+      "learning_rate": 9.952166110886394e-06,
+      "loss": 0.8301,
+      "step": 190
+    },
+    {
+      "epoch": 0.07700452401578593,
+      "grad_norm": 0.10707233101129532,
+      "learning_rate": 9.943176257655567e-06,
+      "loss": 0.8051,
+      "step": 200
+    },
+    {
+      "epoch": 0.08085475021657522,
+      "grad_norm": 0.07714657485485077,
+      "learning_rate": 9.933417550425952e-06,
+      "loss": 0.8126,
+      "step": 210
+    },
+    {
+      "epoch": 0.08470497641736452,
+      "grad_norm": 0.09029405564069748,
+      "learning_rate": 9.9228915070518e-06,
+      "loss": 0.7986,
+      "step": 220
+    },
+    {
+      "epoch": 0.08855520261815382,
+      "grad_norm": 0.09487172216176987,
+      "learning_rate": 9.911599764737632e-06,
+      "loss": 0.7945,
+      "step": 230
+    },
+    {
+      "epoch": 0.09240542881894312,
+      "grad_norm": 0.09710226207971573,
+      "learning_rate": 9.899544079783606e-06,
+      "loss": 0.7906,
+      "step": 240
+    },
+    {
+      "epoch": 0.0962556550197324,
+      "grad_norm": 0.07927985489368439,
+      "learning_rate": 9.886726327312338e-06,
+      "loss": 0.7943,
+      "step": 250
+    },
+    {
+      "epoch": 0.1001058812205217,
+      "grad_norm": 0.0881880521774292,
+      "learning_rate": 9.87314850097724e-06,
+      "loss": 0.8037,
+      "step": 260
+    },
+    {
+      "epoch": 0.103956107421311,
+      "grad_norm": 0.08384250104427338,
+      "learning_rate": 9.858812712652452e-06,
+      "loss": 0.8002,
+      "step": 270
+    },
+    {
+      "epoch": 0.10780633362210029,
+      "grad_norm": 0.09061550348997116,
+      "learning_rate": 9.843721192104341e-06,
+      "loss": 0.7919,
+      "step": 280
+    },
+    {
+      "epoch": 0.11165655982288959,
+      "grad_norm": 0.08281318843364716,
+      "learning_rate": 9.827876286644698e-06,
+      "loss": 0.8062,
+      "step": 290
+    },
+    {
+      "epoch": 0.11550678602367889,
+      "grad_norm": 0.07506729662418365,
+      "learning_rate": 9.811280460765643e-06,
+      "loss": 0.7766,
+      "step": 300
+    },
+    {
+      "epoch": 0.11935701222446819,
+      "grad_norm": 0.08716526627540588,
+      "learning_rate": 9.793936295756292e-06,
+      "loss": 0.7879,
+      "step": 310
+    },
+    {
+      "epoch": 0.12320723842525748,
+      "grad_norm": 0.0750565156340599,
+      "learning_rate": 9.775846489301279e-06,
+      "loss": 0.786,
+      "step": 320
+    },
+    {
+      "epoch": 0.12705746462604678,
+      "grad_norm": 0.08030035346746445,
+      "learning_rate": 9.757013855061157e-06,
+      "loss": 0.7884,
+      "step": 330
+    },
+    {
+      "epoch": 0.13090769082683607,
+      "grad_norm": 0.08966710418462753,
+      "learning_rate": 9.737441322234758e-06,
+      "loss": 0.7989,
+      "step": 340
+    },
+    {
+      "epoch": 0.13475791702762538,
+      "grad_norm": 0.08672546595335007,
+      "learning_rate": 9.717131935103605e-06,
+      "loss": 0.795,
+      "step": 350
+    },
+    {
+      "epoch": 0.13860814322841467,
+      "grad_norm": 0.07555510848760605,
+      "learning_rate": 9.696088852558398e-06,
+      "loss": 0.7813,
+      "step": 360
+    },
+    {
+      "epoch": 0.14245836942920398,
+      "grad_norm": 0.10109313577413559,
+      "learning_rate": 9.674315347607686e-06,
+      "loss": 0.8156,
+      "step": 370
+    },
+    {
+      "epoch": 0.14630859562999327,
+      "grad_norm": 0.07832839339971542,
+      "learning_rate": 9.651814806868796e-06,
+      "loss": 0.7717,
+      "step": 380
+    },
+    {
+      "epoch": 0.15015882183078255,
+      "grad_norm": 0.07502088695764542,
+      "learning_rate": 9.62859073004107e-06,
+      "loss": 0.7863,
+      "step": 390
+    },
+    {
+      "epoch": 0.15400904803157187,
+      "grad_norm": 0.07724788039922714,
+      "learning_rate": 9.604646729361547e-06,
+      "loss": 0.7833,
+      "step": 400
+    },
+    {
+      "epoch": 0.15785927423236115,
+      "grad_norm": 0.07213339954614639,
+      "learning_rate": 9.579986529043099e-06,
+      "loss": 0.7785,
+      "step": 410
+    },
+    {
+      "epoch": 0.16170950043315044,
+      "grad_norm": 0.08125048875808716,
+      "learning_rate": 9.554613964695189e-06,
+      "loss": 0.7758,
+      "step": 420
+    },
+    {
+      "epoch": 0.16555972663393975,
+      "grad_norm": 0.08028566092252731,
+      "learning_rate": 9.528532982727281e-06,
+      "loss": 0.7757,
+      "step": 430
+    },
+    {
+      "epoch": 0.16940995283472904,
+      "grad_norm": 0.10370927304029465,
+      "learning_rate": 9.501747639735018e-06,
+      "loss": 0.7872,
+      "step": 440
+    },
+    {
+      "epoch": 0.17326017903551832,
+      "grad_norm": 0.07269582897424698,
+      "learning_rate": 9.474262101869272e-06,
+      "loss": 0.7595,
+      "step": 450
+    },
+    {
+      "epoch": 0.17711040523630764,
+      "grad_norm": 0.08683127164840698,
+      "learning_rate": 9.446080644188141e-06,
+      "loss": 0.7781,
+      "step": 460
+    },
+    {
+      "epoch": 0.18096063143709692,
+      "grad_norm": 0.08483637124300003,
+      "learning_rate": 9.417207649992017e-06,
+      "loss": 0.7909,
+      "step": 470
+    },
+    {
+      "epoch": 0.18481085763788624,
+      "grad_norm": 0.07935497164726257,
+      "learning_rate": 9.387647610141816e-06,
+      "loss": 0.7736,
+      "step": 480
+    },
+    {
+      "epoch": 0.18866108383867553,
+      "grad_norm": 0.07130106538534164,
+      "learning_rate": 9.35740512236047e-06,
+      "loss": 0.7758,
+      "step": 490
+    },
+    {
+      "epoch": 0.1925113100394648,
+      "grad_norm": 0.06958535313606262,
+      "learning_rate": 9.326484890517809e-06,
+      "loss": 0.7742,
+      "step": 500
+    },
+    {
+      "epoch": 0.19636153624025413,
+      "grad_norm": 0.08078718930482864,
+      "learning_rate": 9.294891723898926e-06,
+      "loss": 0.751,
+      "step": 510
+    },
+    {
+      "epoch": 0.2002117624410434,
+      "grad_norm": 0.07929139584302902,
+      "learning_rate": 9.262630536456153e-06,
+      "loss": 0.7961,
+      "step": 520
+    },
+    {
+      "epoch": 0.2040619886418327,
+      "grad_norm": 0.08046405017375946,
+      "learning_rate": 9.229706346044749e-06,
+      "loss": 0.7566,
+      "step": 530
+    },
+    {
+      "epoch": 0.207912214842622,
+      "grad_norm": 0.08585936576128006,
+      "learning_rate": 9.196124273642422e-06,
+      "loss": 0.7777,
+      "step": 540
+    },
+    {
+      "epoch": 0.2117624410434113,
+      "grad_norm": 0.07530266791582108,
+      "learning_rate": 9.161889542552842e-06,
+      "loss": 0.7833,
+      "step": 550
+    },
+    {
+      "epoch": 0.21561266724420058,
+      "grad_norm": 0.08337376266717911,
+      "learning_rate": 9.127007477593193e-06,
+      "loss": 0.7834,
+      "step": 560
+    },
+    {
+      "epoch": 0.2194628934449899,
+      "grad_norm": 0.08559072762727737,
+      "learning_rate": 9.091483504265974e-06,
+      "loss": 0.7992,
+      "step": 570
+    },
+    {
+      "epoch": 0.22331311964577918,
+      "grad_norm": 0.07006446272134781,
+      "learning_rate": 9.055323147915116e-06,
+      "loss": 0.7682,
+      "step": 580
+    },
+    {
+      "epoch": 0.2271633458465685,
+      "grad_norm": 0.07915423065423965,
+      "learning_rate": 9.018532032866585e-06,
+      "loss": 0.794,
+      "step": 590
+    },
+    {
+      "epoch": 0.23101357204735778,
+      "grad_norm": 0.08861169219017029,
+      "learning_rate": 8.981115881553579e-06,
+      "loss": 0.7864,
+      "step": 600
+    },
+    {
+      "epoch": 0.23486379824814707,
+      "grad_norm": 0.07840580493211746,
+      "learning_rate": 8.943080513626476e-06,
+      "loss": 0.8017,
+      "step": 610
+    },
+    {
+      "epoch": 0.23871402444893638,
+      "grad_norm": 0.07716800272464752,
+      "learning_rate": 8.904431845047641e-06,
+      "loss": 0.7969,
+      "step": 620
+    },
+    {
+      "epoch": 0.24256425064972567,
+      "grad_norm": 0.08346349745988846,
+      "learning_rate": 8.86517588717128e-06,
+      "loss": 0.7915,
+      "step": 630
+    },
+    {
+      "epoch": 0.24641447685051496,
+      "grad_norm": 0.07674357295036316,
+      "learning_rate": 8.82531874580844e-06,
+      "loss": 0.7856,
+      "step": 640
+    },
+    {
+      "epoch": 0.25026470305130427,
+      "grad_norm": 0.06935236603021622,
+      "learning_rate": 8.784866620277314e-06,
+      "loss": 0.7514,
+      "step": 650
+    },
+    {
+      "epoch": 0.25411492925209356,
+      "grad_norm": 0.08185010403394699,
+      "learning_rate": 8.743825802439018e-06,
+      "loss": 0.7817,
+      "step": 660
+    },
+    {
+      "epoch": 0.25796515545288284,
+      "grad_norm": 0.08079718053340912,
+      "learning_rate": 8.702202675718956e-06,
+      "loss": 0.7839,
+      "step": 670
+    },
+    {
+      "epoch": 0.26181538165367213,
+      "grad_norm": 0.07265397161245346,
+      "learning_rate": 8.660003714113959e-06,
+      "loss": 0.7575,
+      "step": 680
+    },
+    {
+      "epoch": 0.26566560785446147,
+      "grad_norm": 0.07351795583963394,
+      "learning_rate": 8.617235481185323e-06,
+      "loss": 0.7914,
+      "step": 690
+    },
+    {
+      "epoch": 0.26951583405525076,
+      "grad_norm": 0.07679812610149384,
+      "learning_rate": 8.573904629037931e-06,
+      "loss": 0.7809,
+      "step": 700
+    },
+    {
+      "epoch": 0.27336606025604004,
+      "grad_norm": 0.07201540470123291,
+      "learning_rate": 8.530017897285585e-06,
+      "loss": 0.7732,
+      "step": 710
+    },
+    {
+      "epoch": 0.27721628645682933,
+      "grad_norm": 0.07132424414157867,
+      "learning_rate": 8.485582112002746e-06,
+      "loss": 0.7709,
+      "step": 720
+    },
+    {
+      "epoch": 0.2810665126576186,
+      "grad_norm": 0.07454054802656174,
+      "learning_rate": 8.440604184662811e-06,
+      "loss": 0.7824,
+      "step": 730
+    },
+    {
+      "epoch": 0.28491673885840796,
+      "grad_norm": 0.08672741800546646,
+      "learning_rate": 8.395091111063118e-06,
+      "loss": 0.7614,
+      "step": 740
+    },
+    {
+      "epoch": 0.28876696505919724,
+      "grad_norm": 0.07655325531959534,
+      "learning_rate": 8.349049970236822e-06,
+      "loss": 0.7675,
+      "step": 750
+    },
+    {
+      "epoch": 0.29261719125998653,
+      "grad_norm": 0.08207186311483383,
+      "learning_rate": 8.302487923351842e-06,
+      "loss": 0.7599,
+      "step": 760
+    },
+    {
+      "epoch": 0.2964674174607758,
+      "grad_norm": 0.08197974413633347,
+      "learning_rate": 8.255412212597021e-06,
+      "loss": 0.7736,
+      "step": 770
+    },
+    {
+      "epoch": 0.3003176436615651,
+      "grad_norm": 0.08144683390855789,
+      "learning_rate": 8.207830160055688e-06,
+      "loss": 0.7642,
+      "step": 780
+    },
+    {
+      "epoch": 0.3041678698623544,
+      "grad_norm": 0.07666800916194916,
+      "learning_rate": 8.159749166566786e-06,
+      "loss": 0.763,
+      "step": 790
+    },
+    {
+      "epoch": 0.30801809606314373,
+      "grad_norm": 0.07657331973314285,
+      "learning_rate": 8.11117671057377e-06,
+      "loss": 0.7817,
+      "step": 800
+    },
+    {
+      "epoch": 0.311868322263933,
+      "grad_norm": 0.0759219080209732,
+      "learning_rate": 8.062120346961404e-06,
+      "loss": 0.7656,
+      "step": 810
+    },
+    {
+      "epoch": 0.3157185484647223,
+      "grad_norm": 0.07638432085514069,
+      "learning_rate": 8.012587705880705e-06,
+      "loss": 0.7751,
+      "step": 820
+    },
+    {
+      "epoch": 0.3195687746655116,
+      "grad_norm": 0.07504768669605255,
+      "learning_rate": 7.962586491562146e-06,
+      "loss": 0.7615,
+      "step": 830
+    },
+    {
+      "epoch": 0.3234190008663009,
+      "grad_norm": 0.07947075366973877,
+      "learning_rate": 7.912124481117357e-06,
+      "loss": 0.7492,
+      "step": 840
+    },
+    {
+      "epoch": 0.3272692270670902,
+      "grad_norm": 0.09768472611904144,
+      "learning_rate": 7.861209523329494e-06,
+      "loss": 0.7695,
+      "step": 850
+    },
+    {
+      "epoch": 0.3311194532678795,
+      "grad_norm": 0.0744275376200676,
+      "learning_rate": 7.809849537432432e-06,
+      "loss": 0.7557,
+      "step": 860
+    },
+    {
+      "epoch": 0.3349696794686688,
+      "grad_norm": 0.07044808566570282,
+      "learning_rate": 7.758052511879035e-06,
+      "loss": 0.756,
+      "step": 870
+    },
+    {
+      "epoch": 0.3388199056694581,
+      "grad_norm": 0.0669751986861229,
+      "learning_rate": 7.705826503098642e-06,
+      "loss": 0.7612,
+      "step": 880
+    },
+    {
+      "epoch": 0.34267013187024736,
+      "grad_norm": 0.07563166320323944,
+      "learning_rate": 7.653179634243979e-06,
+      "loss": 0.7505,
+      "step": 890
+    },
+    {
+      "epoch": 0.34652035807103665,
+      "grad_norm": 0.06728845089673996,
+      "learning_rate": 7.600120093927696e-06,
+      "loss": 0.769,
+      "step": 900
+    },
+    {
+      "epoch": 0.350370584271826,
+      "grad_norm": 0.07853127270936966,
+      "learning_rate": 7.546656134948729e-06,
+      "loss": 0.7529,
+      "step": 910
+    },
+    {
+      "epoch": 0.3542208104726153,
+      "grad_norm": 0.07399027794599533,
+      "learning_rate": 7.492796073008661e-06,
+      "loss": 0.7462,
+      "step": 920
+    },
+    {
+      "epoch": 0.35807103667340456,
+      "grad_norm": 0.08728384226560593,
+      "learning_rate": 7.438548285418326e-06,
+      "loss": 0.7464,
+      "step": 930
+    },
+    {
+      "epoch": 0.36192126287419385,
+      "grad_norm": 0.08288568258285522,
+      "learning_rate": 7.3839212097948e-06,
+      "loss": 0.7697,
+      "step": 940
+    },
+    {
+      "epoch": 0.36577148907498314,
+      "grad_norm": 0.06970415264368057,
+      "learning_rate": 7.328923342749039e-06,
+      "loss": 0.7785,
+      "step": 950
+    },
+    {
+      "epoch": 0.3696217152757725,
+      "grad_norm": 0.08311513811349869,
+      "learning_rate": 7.273563238564319e-06,
+      "loss": 0.7599,
+      "step": 960
+    },
+    {
+      "epoch": 0.37347194147656176,
+      "grad_norm": 0.07381438463926315,
+      "learning_rate": 7.217849507865724e-06,
+      "loss": 0.7504,
+      "step": 970
+    },
+    {
+      "epoch": 0.37732216767735105,
+      "grad_norm": 0.0711706206202507,
+      "learning_rate": 7.161790816280858e-06,
+      "loss": 0.7535,
+      "step": 980
+    },
+    {
+      "epoch": 0.38117239387814034,
+      "grad_norm": 0.08199508488178253,
+      "learning_rate": 7.1053958830920125e-06,
+      "loss": 0.7506,
+      "step": 990
+    },
+    {
+      "epoch": 0.3850226200789296,
+      "grad_norm": 0.07574780285358429,
+      "learning_rate": 7.0486734798799715e-06,
+      "loss": 0.7654,
+      "step": 1000
+    },
+    {
+      "epoch": 0.3888728462797189,
+      "grad_norm": 0.07989233732223511,
+      "learning_rate": 6.991632429159708e-06,
+      "loss": 0.7726,
+      "step": 1010
+    },
+    {
+      "epoch": 0.39272307248050825,
+      "grad_norm": 0.07846470177173615,
+      "learning_rate": 6.934281603008132e-06,
+      "loss": 0.7676,
+      "step": 1020
+    },
+    {
+      "epoch": 0.39657329868129754,
+      "grad_norm": 0.07596123218536377,
+      "learning_rate": 6.876629921684143e-06,
+      "loss": 0.7719,
+      "step": 1030
+    },
+    {
+      "epoch": 0.4004235248820868,
+      "grad_norm": 0.07534731179475784,
+      "learning_rate": 6.818686352241199e-06,
+      "loss": 0.7717,
+      "step": 1040
+    },
+    {
+      "epoch": 0.4042737510828761,
+      "grad_norm": 0.0789870172739029,
+      "learning_rate": 6.760459907132573e-06,
+      "loss": 0.7693,
+      "step": 1050
+    },
+    {
+      "epoch": 0.4081239772836654,
+      "grad_norm": 0.06968405842781067,
+      "learning_rate": 6.701959642809589e-06,
+      "loss": 0.7607,
+      "step": 1060
+    },
+    {
+      "epoch": 0.41197420348445474,
+      "grad_norm": 0.0771174356341362,
+      "learning_rate": 6.643194658312987e-06,
+      "loss": 0.745,
+      "step": 1070
+    },
+    {
+      "epoch": 0.415824429685244,
+      "grad_norm": 0.07416843622922897,
+      "learning_rate": 6.584174093857676e-06,
+      "loss": 0.7843,
+      "step": 1080
+    },
+    {
+      "epoch": 0.4196746558860333,
+      "grad_norm": 0.08142000436782837,
+      "learning_rate": 6.5249071294110735e-06,
+      "loss": 0.7519,
+      "step": 1090
+    },
+    {
+      "epoch": 0.4235248820868226,
+      "grad_norm": 0.07609821856021881,
+      "learning_rate": 6.465402983265274e-06,
+      "loss": 0.7805,
+      "step": 1100
+    },
+    {
+      "epoch": 0.4273751082876119,
+      "grad_norm": 0.07568683475255966,
+      "learning_rate": 6.405670910603243e-06,
+      "loss": 0.7718,
+      "step": 1110
+    },
+    {
+      "epoch": 0.43122533448840117,
+      "grad_norm": 0.09129253029823303,
+      "learning_rate": 6.345720202059291e-06,
+      "loss": 0.7666,
+      "step": 1120
+    },
+    {
+      "epoch": 0.4350755606891905,
+      "grad_norm": 0.07743094116449356,
+      "learning_rate": 6.285560182274008e-06,
+      "loss": 0.7662,
+      "step": 1130
+    },
+    {
+      "epoch": 0.4389257868899798,
+      "grad_norm": 0.07708477228879929,
+      "learning_rate": 6.225200208443936e-06,
+      "loss": 0.7556,
+      "step": 1140
+    },
+    {
+      "epoch": 0.4427760130907691,
+      "grad_norm": 0.07237216085195541,
+      "learning_rate": 6.16464966886616e-06,
+      "loss": 0.7617,
+      "step": 1150
+    },
+    {
+      "epoch": 0.44662623929155837,
+      "grad_norm": 0.07941339164972305,
+      "learning_rate": 6.103917981478065e-06,
+      "loss": 0.7787,
+      "step": 1160
+    },
+    {
+      "epoch": 0.45047646549234766,
+      "grad_norm": 0.07328004390001297,
+      "learning_rate": 6.043014592392489e-06,
+      "loss": 0.7861,
+      "step": 1170
+    },
+    {
+      "epoch": 0.454326691693137,
+      "grad_norm": 0.07648901641368866,
+      "learning_rate": 5.981948974428485e-06,
+      "loss": 0.7576,
+      "step": 1180
+    },
+    {
+      "epoch": 0.4581769178939263,
+      "grad_norm": 0.07756921648979187,
+      "learning_rate": 5.920730625637934e-06,
+      "loss": 0.7416,
+      "step": 1190
+    },
+    {
+      "epoch": 0.46202714409471557,
+      "grad_norm": 0.07935123890638351,
+      "learning_rate": 5.8593690678282355e-06,
+      "loss": 0.7535,
+      "step": 1200
+    },
+    {
+      "epoch": 0.46587737029550486,
+      "grad_norm": 0.08051187545061111,
+      "learning_rate": 5.797873845081301e-06,
+      "loss": 0.7536,
+      "step": 1210
+    },
+    {
+      "epoch": 0.46972759649629414,
+      "grad_norm": 0.07447057217359543,
+      "learning_rate": 5.736254522269075e-06,
+      "loss": 0.7765,
+      "step": 1220
+    },
+    {
+      "epoch": 0.47357782269708343,
+      "grad_norm": 0.07386547327041626,
+      "learning_rate": 5.6745206835658386e-06,
+      "loss": 0.7659,
+      "step": 1230
+    },
+    {
+      "epoch": 0.47742804889787277,
+      "grad_norm": 0.06856486946344376,
+      "learning_rate": 5.612681930957495e-06,
+      "loss": 0.7493,
+      "step": 1240
+    },
+    {
+      "epoch": 0.48127827509866206,
+      "grad_norm": 0.07254555821418762,
+      "learning_rate": 5.550747882748095e-06,
+      "loss": 0.7481,
+      "step": 1250
+    },
+    {
+      "epoch": 0.48512850129945134,
+      "grad_norm": 0.07836291939020157,
+      "learning_rate": 5.4887281720638195e-06,
+      "loss": 0.7524,
+      "step": 1260
+    },
+    {
+      "epoch": 0.48897872750024063,
+      "grad_norm": 0.07895445078611374,
+      "learning_rate": 5.42663244535466e-06,
+      "loss": 0.7395,
+      "step": 1270
+    },
+    {
+      "epoch": 0.4928289537010299,
+      "grad_norm": 0.07210540771484375,
+      "learning_rate": 5.3644703608940205e-06,
+      "loss": 0.7488,
+      "step": 1280
+    },
+    {
+      "epoch": 0.49667917990181926,
+      "grad_norm": 0.0756601020693779,
+      "learning_rate": 5.302251587276491e-06,
+      "loss": 0.7738,
+      "step": 1290
+    },
+    {
+      "epoch": 0.5005294061026085,
+      "grad_norm": 0.0745985358953476,
+      "learning_rate": 5.2399858019140005e-06,
+      "loss": 0.7603,
+      "step": 1300
+    },
+    {
+      "epoch": 0.5043796323033978,
+      "grad_norm": 0.07277224212884903,
+      "learning_rate": 5.177682689530615e-06,
+      "loss": 0.7688,
+      "step": 1310
+    },
+    {
+      "epoch": 0.5082298585041871,
+      "grad_norm": 0.07988592982292175,
+      "learning_rate": 5.11535194065619e-06,
+      "loss": 0.7574,
+      "step": 1320
+    },
+    {
+      "epoch": 0.5120800847049765,
+      "grad_norm": 0.09061582386493683,
+      "learning_rate": 5.053003250119113e-06,
+      "loss": 0.7562,
+      "step": 1330
+    },
+    {
+      "epoch": 0.5159303109057657,
+      "grad_norm": 0.07673893868923187,
+      "learning_rate": 4.9906463155383925e-06,
+      "loss": 0.7621,
+      "step": 1340
+    },
+    {
+      "epoch": 0.519780537106555,
+      "grad_norm": 0.08025459945201874,
+      "learning_rate": 4.9282908358153046e-06,
+      "loss": 0.7486,
+      "step": 1350
+    },
+    {
+      "epoch": 0.5236307633073443,
+      "grad_norm": 0.0847226232290268,
+      "learning_rate": 4.8659465096248345e-06,
+      "loss": 0.7627,
+      "step": 1360
+    },
+    {
+      "epoch": 0.5274809895081336,
+      "grad_norm": 0.07819210737943649,
+      "learning_rate": 4.803623033907169e-06,
+      "loss": 0.7649,
+      "step": 1370
+    },
+    {
+      "epoch": 0.5313312157089229,
+      "grad_norm": 0.07083548605442047,
+      "learning_rate": 4.741330102359441e-06,
+      "loss": 0.7768,
+      "step": 1380
+    },
+    {
+      "epoch": 0.5351814419097122,
+      "grad_norm": 0.07928690314292908,
+      "learning_rate": 4.679077403927996e-06,
+      "loss": 0.7471,
+      "step": 1390
+    },
+    {
+      "epoch": 0.5390316681105015,
+      "grad_norm": 0.07777705043554306,
+      "learning_rate": 4.616874621301373e-06,
+      "loss": 0.7504,
+      "step": 1400
+    },
+    {
+      "epoch": 0.5428818943112907,
+      "grad_norm": 0.08884922415018082,
+      "learning_rate": 4.554731429404293e-06,
+      "loss": 0.7738,
+      "step": 1410
+    },
+    {
+      "epoch": 0.5467321205120801,
+      "grad_norm": 0.08875343948602676,
+      "learning_rate": 4.492657493892822e-06,
+      "loss": 0.7561,
+      "step": 1420
+    },
+    {
+      "epoch": 0.5505823467128694,
+      "grad_norm": 0.08747640997171402,
+      "learning_rate": 4.430662469650999e-06,
+      "loss": 0.7604,
+      "step": 1430
+    },
+    {
+      "epoch": 0.5544325729136587,
+      "grad_norm": 0.07842380553483963,
+      "learning_rate": 4.368755999289118e-06,
+      "loss": 0.774,
+      "step": 1440
+    },
+    {
+      "epoch": 0.558282799114448,
+      "grad_norm": 0.08660931140184402,
+      "learning_rate": 4.306947711643949e-06,
+      "loss": 0.7537,
+      "step": 1450
+    },
+    {
+      "epoch": 0.5621330253152372,
+      "grad_norm": 0.07852008938789368,
+      "learning_rate": 4.245247220281068e-06,
+      "loss": 0.7571,
+      "step": 1460
+    },
+    {
+      "epoch": 0.5659832515160266,
+      "grad_norm": 0.07974742352962494,
+      "learning_rate": 4.183664121999587e-06,
+      "loss": 0.7509,
+      "step": 1470
+    },
+    {
+      "epoch": 0.5698334777168159,
+      "grad_norm": 0.07593739032745361,
+      "learning_rate": 4.122207995339474e-06,
+      "loss": 0.7476,
+      "step": 1480
+    },
+    {
+      "epoch": 0.5736837039176051,
+      "grad_norm": 0.08665842562913895,
+      "learning_rate": 4.060888399091734e-06,
+      "loss": 0.7551,
+      "step": 1490
+    },
+    {
+      "epoch": 0.5775339301183945,
+      "grad_norm": 0.0718807503581047,
+      "learning_rate": 3.999714870811639e-06,
+      "loss": 0.7688,
+      "step": 1500
+    },
+    {
+      "epoch": 0.5813841563191837,
+      "grad_norm": 0.07715797424316406,
+      "learning_rate": 3.938696925335275e-06,
+      "loss": 0.7682,
+      "step": 1510
+    },
+    {
+      "epoch": 0.5852343825199731,
+      "grad_norm": 0.07482390850782394,
+      "learning_rate": 3.87784405329962e-06,
+      "loss": 0.7643,
+      "step": 1520
+    },
+    {
+      "epoch": 0.5890846087207623,
+      "grad_norm": 0.07503031194210052,
+      "learning_rate": 3.817165719666382e-06,
+      "loss": 0.7723,
+      "step": 1530
+    },
+    {
+      "epoch": 0.5929348349215516,
+      "grad_norm": 0.07083340734243393,
+      "learning_rate": 3.756671362249841e-06,
+      "loss": 0.7667,
+      "step": 1540
+    },
+    {
+      "epoch": 0.596785061122341,
+      "grad_norm": 0.07452557235956192,
+      "learning_rate": 3.6963703902488907e-06,
+      "loss": 0.7723,
+      "step": 1550
+    },
+    {
+      "epoch": 0.6006352873231302,
+      "grad_norm": 0.0817946046590805,
+      "learning_rate": 3.6362721827835655e-06,
+      "loss": 0.7767,
+      "step": 1560
+    },
+    {
+      "epoch": 0.6044855135239195,
+      "grad_norm": 0.07772989571094513,
+      "learning_rate": 3.576386087436212e-06,
+      "loss": 0.7593,
+      "step": 1570
+    },
+    {
+      "epoch": 0.6083357397247088,
+      "grad_norm": 0.08103422075510025,
+      "learning_rate": 3.5167214187975913e-06,
+      "loss": 0.7636,
+      "step": 1580
+    },
+    {
+      "epoch": 0.6121859659254981,
+      "grad_norm": 0.0697210505604744,
+      "learning_rate": 3.457287457018086e-06,
+      "loss": 0.7714,
+      "step": 1590
+    },
+    {
+      "epoch": 0.6160361921262875,
+      "grad_norm": 0.08649379760026932,
+      "learning_rate": 3.3980934463643003e-06,
+      "loss": 0.7545,
+      "step": 1600
+    },
+    {
+      "epoch": 0.6198864183270767,
+      "grad_norm": 0.07391930371522903,
+      "learning_rate": 3.3391485937812053e-06,
+      "loss": 0.7324,
+      "step": 1610
+    },
+    {
+      "epoch": 0.623736644527866,
+      "grad_norm": 0.07247166335582733,
+      "learning_rate": 3.280462067460115e-06,
+      "loss": 0.7325,
+      "step": 1620
+    },
+    {
+      "epoch": 0.6275868707286553,
+      "grad_norm": 0.07459379732608795,
+      "learning_rate": 3.222042995412669e-06,
+      "loss": 0.7646,
+      "step": 1630
+    },
+    {
+      "epoch": 0.6314370969294446,
+      "grad_norm": 0.06901247054338455,
+      "learning_rate": 3.163900464051089e-06,
+      "loss": 0.7514,
+      "step": 1640
+    },
+    {
+      "epoch": 0.635287323130234,
+      "grad_norm": 0.07590402662754059,
+      "learning_rate": 3.1060435167748826e-06,
+      "loss": 0.7604,
+      "step": 1650
+    },
+    {
+      "epoch": 0.6391375493310232,
+      "grad_norm": 0.07103829085826874,
+      "learning_rate": 3.048481152564253e-06,
+      "loss": 0.7415,
+      "step": 1660
+    },
+    {
+      "epoch": 0.6429877755318125,
+      "grad_norm": 0.08009158819913864,
+      "learning_rate": 2.9912223245804017e-06,
+      "loss": 0.7656,
+      "step": 1670
+    },
+    {
+      "epoch": 0.6468380017326018,
+      "grad_norm": 0.07657431066036224,
+      "learning_rate": 2.934275938772981e-06,
+      "loss": 0.7552,
+      "step": 1680
+    },
+    {
+      "epoch": 0.6506882279333911,
+      "grad_norm": 0.06848358362913132,
+      "learning_rate": 2.8776508524948676e-06,
+      "loss": 0.7475,
+      "step": 1690
+    },
+    {
+      "epoch": 0.6545384541341804,
+      "grad_norm": 0.07067371159791946,
+      "learning_rate": 2.821355873124503e-06,
+      "loss": 0.7491,
+      "step": 1700
+    },
+    {
+      "epoch": 0.6583886803349697,
+      "grad_norm": 0.0827857255935669,
+      "learning_rate": 2.7653997566960144e-06,
+      "loss": 0.7631,
+      "step": 1710
+    },
+    {
+      "epoch": 0.662238906535759,
+      "grad_norm": 0.07938899099826813,
+      "learning_rate": 2.709791206537309e-06,
+      "loss": 0.7519,
+      "step": 1720
+    },
+    {
+      "epoch": 0.6660891327365482,
+      "grad_norm": 0.07389482855796814,
+      "learning_rate": 2.6545388719163747e-06,
+      "loss": 0.7579,
+      "step": 1730
+    },
+    {
+      "epoch": 0.6699393589373376,
+      "grad_norm": 0.08058244735002518,
+      "learning_rate": 2.599651346695979e-06,
+      "loss": 0.7577,
+      "step": 1740
+    },
+    {
+      "epoch": 0.6737895851381268,
+      "grad_norm": 0.07189187407493591,
+      "learning_rate": 2.545137167997007e-06,
+      "loss": 0.7612,
+      "step": 1750
+    },
+    {
+      "epoch": 0.6776398113389162,
+      "grad_norm": 0.07711486518383026,
+      "learning_rate": 2.4910048148705972e-06,
+      "loss": 0.7497,
+      "step": 1760
+    },
+    {
+      "epoch": 0.6814900375397055,
+      "grad_norm": 0.0764533281326294,
+      "learning_rate": 2.4372627069793327e-06,
+      "loss": 0.78,
+      "step": 1770
+    },
+    {
+      "epoch": 0.6853402637404947,
+      "grad_norm": 0.07591167092323303,
+      "learning_rate": 2.3839192032876486e-06,
+      "loss": 0.7691,
+      "step": 1780
+    },
+    {
+      "epoch": 0.6891904899412841,
+      "grad_norm": 0.07528550177812576,
+      "learning_rate": 2.3309826007617047e-06,
+      "loss": 0.7813,
+      "step": 1790
+    },
+    {
+      "epoch": 0.6930407161420733,
+      "grad_norm": 0.06685158610343933,
+      "learning_rate": 2.278461133078881e-06,
+      "loss": 0.7438,
+      "step": 1800
+    },
+    {
+      "epoch": 0.6968909423428626,
+      "grad_norm": 0.0733516663312912,
+      "learning_rate": 2.2263629693471258e-06,
+      "loss": 0.7496,
+      "step": 1810
+    },
+    {
+      "epoch": 0.700741168543652,
+      "grad_norm": 0.07055472582578659,
+      "learning_rate": 2.1746962128343414e-06,
+      "loss": 0.7692,
+      "step": 1820
+    },
+    {
+      "epoch": 0.7045913947444412,
+      "grad_norm": 0.07128188014030457,
+      "learning_rate": 2.1234688997080183e-06,
+      "loss": 0.7614,
+      "step": 1830
+    },
+    {
+      "epoch": 0.7084416209452306,
+      "grad_norm": 0.08089336007833481,
+      "learning_rate": 2.0726889977853053e-06,
+      "loss": 0.7569,
+      "step": 1840
+    },
+    {
+      "epoch": 0.7122918471460198,
+      "grad_norm": 0.07378948479890823,
+      "learning_rate": 2.022364405293703e-06,
+      "loss": 0.7612,
+      "step": 1850
+    },
+    {
+      "epoch": 0.7161420733468091,
+      "grad_norm": 0.07310800999403,
+      "learning_rate": 1.9725029496425814e-06,
+      "loss": 0.7684,
+      "step": 1860
+    },
+    {
+      "epoch": 0.7199922995475985,
+      "grad_norm": 0.07425514608621597,
+      "learning_rate": 1.923112386205727e-06,
+      "loss": 0.76,
+      "step": 1870
+    },
+    {
+      "epoch": 0.7238425257483877,
+      "grad_norm": 0.072078637778759,
+      "learning_rate": 1.874200397115079e-06,
+      "loss": 0.764,
+      "step": 1880
+    },
+    {
+      "epoch": 0.727692751949177,
+      "grad_norm": 0.07185759395360947,
+      "learning_rate": 1.825774590065857e-06,
+      "loss": 0.7384,
+      "step": 1890
+    },
+    {
+      "epoch": 0.7315429781499663,
+      "grad_norm": 0.07672592997550964,
+      "learning_rate": 1.7778424971332813e-06,
+      "loss": 0.7652,
+      "step": 1900
+    },
+    {
+      "epoch": 0.7353932043507556,
+      "grad_norm": 0.07018318772315979,
+      "learning_rate": 1.7304115736010407e-06,
+      "loss": 0.7596,
+      "step": 1910
+    },
+    {
+      "epoch": 0.739243430551545,
+      "grad_norm": 0.07700502127408981,
+      "learning_rate": 1.6834891968017108e-06,
+      "loss": 0.7433,
+      "step": 1920
+    },
+    {
+      "epoch": 0.7430936567523342,
+      "grad_norm": 0.08042623847723007,
+      "learning_rate": 1.6370826649692884e-06,
+      "loss": 0.7677,
+      "step": 1930
+    },
+    {
+      "epoch": 0.7469438829531235,
+      "grad_norm": 0.07124733179807663,
+      "learning_rate": 1.5911991961040463e-06,
+      "loss": 0.7663,
+      "step": 1940
+    },
+    {
+      "epoch": 0.7507941091539128,
+      "grad_norm": 0.07197007536888123,
+      "learning_rate": 1.545845926849851e-06,
+      "loss": 0.7614,
+      "step": 1950
+    },
+    {
+      "epoch": 0.7546443353547021,
+      "grad_norm": 0.07262544333934784,
+      "learning_rate": 1.5010299113841397e-06,
+      "loss": 0.782,
+      "step": 1960
+    },
+    {
+      "epoch": 0.7584945615554913,
+      "grad_norm": 0.07326176762580872,
+      "learning_rate": 1.4567581203207238e-06,
+      "loss": 0.7619,
+      "step": 1970
+    },
+    {
+      "epoch": 0.7623447877562807,
+      "grad_norm": 0.07612668722867966,
+      "learning_rate": 1.413037439625593e-06,
+      "loss": 0.7557,
+      "step": 1980
+    },
+    {
+      "epoch": 0.76619501395707,
+      "grad_norm": 0.0821150541305542,
+      "learning_rate": 1.3698746695458847e-06,
+      "loss": 0.7685,
+      "step": 1990
+    },
+    {
+      "epoch": 0.7700452401578592,
+      "grad_norm": 0.06652616709470749,
+      "learning_rate": 1.3272765235521745e-06,
+      "loss": 0.7528,
+      "step": 2000
+    },
+    {
+      "epoch": 0.7738954663586486,
+      "grad_norm": 0.07544726878404617,
+      "learning_rate": 1.2852496272942877e-06,
+      "loss": 0.7573,
+      "step": 2010
+    },
+    {
+      "epoch": 0.7777456925594378,
+      "grad_norm": 0.0672341138124466,
+      "learning_rate": 1.2438005175707396e-06,
+      "loss": 0.7485,
+      "step": 2020
+    },
+    {
+      "epoch": 0.7815959187602272,
+      "grad_norm": 0.07542672008275986,
+      "learning_rate": 1.2029356413120297e-06,
+      "loss": 0.7418,
+      "step": 2030
+    },
+    {
+      "epoch": 0.7854461449610165,
+      "grad_norm": 0.07497099786996841,
+      "learning_rate": 1.1626613545778786e-06,
+      "loss": 0.7563,
+      "step": 2040
+    },
+    {
+      "epoch": 0.7892963711618057,
+      "grad_norm": 0.07084732502698898,
+      "learning_rate": 1.1229839215686294e-06,
+      "loss": 0.7572,
+      "step": 2050
+    },
+    {
+      "epoch": 0.7931465973625951,
+      "grad_norm": 0.0699266716837883,
+      "learning_rate": 1.0839095136509188e-06,
+      "loss": 0.7432,
+      "step": 2060
+    },
+    {
+      "epoch": 0.7969968235633843,
+      "grad_norm": 0.07894887030124664,
+      "learning_rate": 1.045444208397791e-06,
+      "loss": 0.7643,
+      "step": 2070
+    },
+    {
+      "epoch": 0.8008470497641736,
+      "grad_norm": 0.06369058042764664,
+      "learning_rate": 1.007593988643402e-06,
+      "loss": 0.7377,
+      "step": 2080
+    },
+    {
+      "epoch": 0.804697275964963,
+      "grad_norm": 0.06588619947433472,
+      "learning_rate": 9.703647415524613e-07,
+      "loss": 0.7611,
+      "step": 2090
+    },
+    {
+      "epoch": 0.8085475021657522,
+      "grad_norm": 0.07450055330991745,
+      "learning_rate": 9.337622577045485e-07,
+      "loss": 0.7593,
+      "step": 2100
+    },
+    {
+      "epoch": 0.8123977283665416,
+      "grad_norm": 0.0743057131767273,
+      "learning_rate": 8.977922301934588e-07,
+      "loss": 0.757,
+      "step": 2110
+    },
+    {
+      "epoch": 0.8162479545673308,
+      "grad_norm": 0.07248660176992416,
+      "learning_rate": 8.624602537417009e-07,
+      "loss": 0.7579,
+      "step": 2120
+    },
+    {
+      "epoch": 0.8200981807681201,
+      "grad_norm": 0.08066260069608688,
+      "learning_rate": 8.277718238303123e-07,
+      "loss": 0.7543,
+      "step": 2130
+    },
+    {
+      "epoch": 0.8239484069689095,
+      "grad_norm": 0.07386893779039383,
+      "learning_rate": 7.937323358440935e-07,
+      "loss": 0.758,
+      "step": 2140
+    },
+    {
+      "epoch": 0.8277986331696987,
+      "grad_norm": 0.08627674728631973,
+      "learning_rate": 7.603470842324223e-07,
+      "loss": 0.752,
+      "step": 2150
+    },
+    {
+      "epoch": 0.831648859370488,
+      "grad_norm": 0.07151111960411072,
+      "learning_rate": 7.276212616857575e-07,
+      "loss": 0.7689,
+      "step": 2160
+    },
+    {
+      "epoch": 0.8354990855712773,
+      "grad_norm": 0.07377808541059494,
+      "learning_rate": 6.955599583279837e-07,
+      "loss": 0.7688,
+      "step": 2170
+    },
+    {
+      "epoch": 0.8393493117720666,
+      "grad_norm": 0.07049500942230225,
+      "learning_rate": 6.641681609246981e-07,
+      "loss": 0.7541,
+      "step": 2180
+    },
+    {
+      "epoch": 0.8431995379728559,
+      "grad_norm": 0.08067857474088669,
+      "learning_rate": 6.334507521075711e-07,
+      "loss": 0.7593,
+      "step": 2190
+    },
+    {
+      "epoch": 0.8470497641736452,
+      "grad_norm": 0.08447853475809097,
+      "learning_rate": 6.034125096149179e-07,
+      "loss": 0.753,
+      "step": 2200
+    },
+    {
+      "epoch": 0.8508999903744345,
+      "grad_norm": 0.08030492812395096,
+      "learning_rate": 5.740581055485722e-07,
+      "loss": 0.7346,
+      "step": 2210
+    },
+    {
+      "epoch": 0.8547502165752238,
+      "grad_norm": 0.07231003791093826,
+      "learning_rate": 5.453921056471961e-07,
+      "loss": 0.7673,
+      "step": 2220
+    },
+    {
+      "epoch": 0.8586004427760131,
+      "grad_norm": 0.0730845257639885,
+      "learning_rate": 5.174189685761288e-07,
+      "loss": 0.7451,
+      "step": 2230
+    },
+    {
+      "epoch": 0.8624506689768023,
+      "grad_norm": 0.07229430228471756,
+      "learning_rate": 4.901430452338979e-07,
+      "loss": 0.7684,
+      "step": 2240
+    },
+    {
+      "epoch": 0.8663008951775917,
+      "grad_norm": 0.07005860656499863,
+      "learning_rate": 4.6356857807548286e-07,
+      "loss": 0.7546,
+      "step": 2250
+    },
+    {
+      "epoch": 0.870151121378381,
+      "grad_norm": 0.07827958464622498,
+      "learning_rate": 4.376997004524541e-07,
+      "loss": 0.748,
+      "step": 2260
+    },
+    {
+      "epoch": 0.8740013475791703,
+      "grad_norm": 0.0767517015337944,
+      "learning_rate": 4.12540435970073e-07,
+      "loss": 0.7568,
+      "step": 2270
+    },
+    {
+      "epoch": 0.8778515737799596,
+      "grad_norm": 0.06849540770053864,
+      "learning_rate": 3.880946978614703e-07,
+      "loss": 0.7582,
+      "step": 2280
+    },
+    {
+      "epoch": 0.8817017999807488,
+      "grad_norm": 0.06945688277482986,
+      "learning_rate": 3.643662883789878e-07,
+      "loss": 0.7701,
+      "step": 2290
+    },
+    {
+      "epoch": 0.8855520261815382,
+      "grad_norm": 0.07245782762765884,
+      "learning_rate": 3.41358898202781e-07,
+      "loss": 0.7476,
+      "step": 2300
+    },
+    {
+      "epoch": 0.8894022523823275,
+      "grad_norm": 0.06896769255399704,
+      "learning_rate": 3.190761058667741e-07,
+      "loss": 0.7573,
+      "step": 2310
+    },
+    {
+      "epoch": 0.8932524785831167,
+      "grad_norm": 0.07035563886165619,
+      "learning_rate": 2.9752137720206463e-07,
+      "loss": 0.7619,
+      "step": 2320
+    },
+    {
+      "epoch": 0.8971027047839061,
+      "grad_norm": 0.07263980060815811,
+      "learning_rate": 2.7669806479785067e-07,
+      "loss": 0.7689,
+      "step": 2330
+    },
+    {
+      "epoch": 0.9009529309846953,
+      "grad_norm": 0.06894528865814209,
+      "learning_rate": 2.566094074799763e-07,
+      "loss": 0.7549,
+      "step": 2340
+    },
+    {
+      "epoch": 0.9048031571854847,
+      "grad_norm": 0.07541150599718094,
+      "learning_rate": 2.3725852980716557e-07,
+      "loss": 0.7399,
+      "step": 2350
+    },
+    {
+      "epoch": 0.908653383386274,
+      "grad_norm": 0.08217129856348038,
+      "learning_rate": 2.1864844158503983e-07,
+      "loss": 0.7608,
+      "step": 2360
+    },
+    {
+      "epoch": 0.9125036095870632,
+      "grad_norm": 0.08027851581573486,
+      "learning_rate": 2.007820373979702e-07,
+      "loss": 0.7402,
+      "step": 2370
+    },
+    {
+      "epoch": 0.9163538357878526,
+      "grad_norm": 0.07157991081476212,
+      "learning_rate": 1.8366209615886532e-07,
+      "loss": 0.7541,
+      "step": 2380
+    },
+    {
+      "epoch": 0.9202040619886418,
+      "grad_norm": 0.07188992947340012,
+      "learning_rate": 1.6729128067693613e-07,
+      "loss": 0.7716,
+      "step": 2390
+    },
+    {
+      "epoch": 0.9240542881894311,
+      "grad_norm": 0.07231468707323074,
+      "learning_rate": 1.5167213724353426e-07,
+      "loss": 0.7558,
+      "step": 2400
+    },
+    {
+      "epoch": 0.9279045143902205,
+      "grad_norm": 0.07128830254077911,
+      "learning_rate": 1.3680709523610058e-07,
+      "loss": 0.7424,
+      "step": 2410
+    },
+    {
+      "epoch": 0.9317547405910097,
+      "grad_norm": 0.0740322694182396,
+      "learning_rate": 1.2269846674030418e-07,
+      "loss": 0.7747,
+      "step": 2420
+    },
+    {
+      "epoch": 0.935604966791799,
+      "grad_norm": 0.07381503283977509,
+      "learning_rate": 1.0934844619042595e-07,
+      "loss": 0.7645,
+      "step": 2430
+    },
+    {
+      "epoch": 0.9394551929925883,
+      "grad_norm": 0.0766066163778305,
+      "learning_rate": 9.675911002803961e-08,
+      "loss": 0.7509,
+      "step": 2440
+    },
+    {
+      "epoch": 0.9433054191933776,
+      "grad_norm": 0.07265421748161316,
+      "learning_rate": 8.493241637904381e-08,
+      "loss": 0.7476,
+      "step": 2450
+    },
+    {
+      "epoch": 0.9471556453941669,
+      "grad_norm": 0.076146699488163,
+      "learning_rate": 7.387020474909689e-08,
+      "loss": 0.7727,
+      "step": 2460
+    },
+    {
+      "epoch": 0.9510058715949562,
+      "grad_norm": 0.07386243343353271,
+      "learning_rate": 6.357419573750856e-08,
+      "loss": 0.7522,
+      "step": 2470
+    },
+    {
+      "epoch": 0.9548560977957455,
+      "grad_norm": 0.07493852078914642,
+      "learning_rate": 5.4045990769613365e-08,
+      "loss": 0.7581,
+      "step": 2480
+    },
+    {
+      "epoch": 0.9587063239965348,
+      "grad_norm": 0.07243316620588303,
+      "learning_rate": 4.5287071847694434e-08,
+      "loss": 0.7556,
+      "step": 2490
+    },
+    {
+      "epoch": 0.9625565501973241,
+      "grad_norm": 0.07479245960712433,
+      "learning_rate": 3.7298801320467817e-08,
+      "loss": 0.746,
+      "step": 2500
+    },
+    {
+      "epoch": 0.9664067763981133,
+      "grad_norm": 0.07309582084417343,
+      "learning_rate": 3.0082421671192576e-08,
+      "loss": 0.7564,
+      "step": 2510
+    },
+    {
+      "epoch": 0.9702570025989027,
+      "grad_norm": 0.06835458427667618,
+      "learning_rate": 2.3639055324410355e-08,
+      "loss": 0.7516,
+      "step": 2520
+    },
+    {
+      "epoch": 0.974107228799692,
+      "grad_norm": 0.08302287757396698,
+      "learning_rate": 1.7969704471369477e-08,
+      "loss": 0.7467,
+      "step": 2530
+    },
+    {
+      "epoch": 0.9779574550004813,
+      "grad_norm": 0.0694880485534668,
+      "learning_rate": 1.3075250914142413e-08,
+      "loss": 0.7531,
+      "step": 2540
+    },
+    {
+      "epoch": 0.9818076812012706,
+      "grad_norm": 0.0708552822470665,
+      "learning_rate": 8.956455928476049e-09,
+      "loss": 0.7659,
+      "step": 2550
+    },
+    {
+      "epoch": 0.9856579074020598,
+      "grad_norm": 0.06500305980443954,
+      "learning_rate": 5.6139601453786275e-09,
+      "loss": 0.7439,
+      "step": 2560
+    },
+    {
+      "epoch": 0.9895081336028492,
+      "grad_norm": 0.07047093659639359,
+      "learning_rate": 3.0482834514805602e-09,
+      "loss": 0.7431,
+      "step": 2570
+    },
+    {
+      "epoch": 0.9933583598036385,
+      "grad_norm": 0.07034512609243393,
+      "learning_rate": 1.2598249081713366e-09,
+      "loss": 0.7338,
+      "step": 2580
+    },
+    {
+      "epoch": 0.9972085860044277,
+      "grad_norm": 0.07195116579532623,
+      "learning_rate": 2.488626895302826e-10,
+      "loss": 0.7624,
+      "step": 2590
+    },
+    {
+      "epoch": 0.9999037443449803,
+      "step": 2597,
+      "total_flos": 2.877231303135068e+18,
+      "train_loss": 0.7796144178843655,
+      "train_runtime": 12744.7978,
+      "train_samples_per_second": 52.171,
+      "train_steps_per_second": 0.204
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 2597,
+  "num_input_tokens_seen": 246462940,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.877231303135068e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

continuous/qwen25/pixio_vith16/finetune/training_args.bin ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:bb4cedc2e8654086a6bd4dd9ef46470cea91141e966be985c1405095d4164090
+size 6264

continuous/qwen25/pixio_vith16/finetune/vocab.json ADDED Viewed

The diff for this file is too large to render. See raw diff

continuous/qwen25/pixio_vith16/pretrain.bad_hf_native_20260818/config.json ADDED Viewed

	@@ -0,0 +1,87 @@

+{
+  "add_faster_video": false,
+  "add_time_instruction": false,
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "dtype": "bfloat16",
+  "eos_token_id": 151645,
+  "faster_token_stride": 10,
+  "force_sample": false,
+  "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,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "mm_hidden_size": 1280,
+  "mm_newline_position": "grid",
+  "mm_patch_merge_type": "flat",
+  "mm_projector_lr": null,
+  "mm_projector_type": "mlp2x_gelu",
+  "mm_resampler_type": null,
+  "mm_spatial_pool_mode": "bilinear",
+  "mm_spatial_pool_stride": null,
+  "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": -1,
+  "mm_vision_tower": "vtb_ssl:pixio",
+  "mm_vision_tower_lr": null,
+  "model_type": "qwen2",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 2,
+  "pad_token_id": 151643,
+  "pos_skipping_range": 4096,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "tokenizer_model_max_length": 8192,
+  "tokenizer_padding_side": "right",
+  "transformers_version": "4.57.6",
+  "use_cache": true,
+  "use_mm_proj": true,
+  "use_pos_skipping": false,
+  "use_sliding_window": false,
+  "vision_tower_pretrained": "/cache/ckpt/download/tokenizer/continuous/pixio/pixio_vith16/model.native.pth",
+  "vocab_size": 151936
+}

continuous/qwen25/pixio_vith16/pretrain.bad_hf_native_20260818/mm_projector.bin ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:9248ee628dfc8dfadc1a1ec160f256937a55e16a94b69ca5b58d4e2f781e5fb5
+size 17315896

continuous/qwen25/pixio_vith16/pretrain.bad_hf_native_20260818/trainer_state.json ADDED Viewed

	@@ -0,0 +1,421 @@

+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 545,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01834862385321101,
+      "grad_norm": 0.13427352905273438,
+      "learning_rate": 0.00032142857142857147,
+      "loss": 4.8135,
+      "step": 10
+    },
+    {
+      "epoch": 0.03669724770642202,
+      "grad_norm": 0.24668659269809723,
+      "learning_rate": 0.0006785714285714287,
+      "loss": 4.1789,
+      "step": 20
+    },
+    {
+      "epoch": 0.05504587155963303,
+      "grad_norm": 0.046047668904066086,
+      "learning_rate": 0.0009999907688176173,
+      "loss": 3.6663,
+      "step": 30
+    },
+    {
+      "epoch": 0.07339449541284404,
+      "grad_norm": 0.026219546794891357,
+      "learning_rate": 0.0009988834393115766,
+      "loss": 3.3958,
+      "step": 40
+    },
+    {
+      "epoch": 0.09174311926605505,
+      "grad_norm": 0.030715452507138252,
+      "learning_rate": 0.0009959345572940772,
+      "loss": 3.3282,
+      "step": 50
+    },
+    {
+      "epoch": 0.11009174311926606,
+      "grad_norm": 0.02611163817346096,
+      "learning_rate": 0.0009911550081155983,
+      "loss": 3.2921,
+      "step": 60
+    },
+    {
+      "epoch": 0.12844036697247707,
+      "grad_norm": 0.014269200153648853,
+      "learning_rate": 0.0009845624347566433,
+      "loss": 3.2729,
+      "step": 70
+    },
+    {
+      "epoch": 0.14678899082568808,
+      "grad_norm": 0.017834672704339027,
+      "learning_rate": 0.0009761811727013547,
+      "loss": 3.2603,
+      "step": 80
+    },
+    {
+      "epoch": 0.1651376146788991,
+      "grad_norm": 0.012764865532517433,
+      "learning_rate": 0.0009660421601067666,
+      "loss": 3.246,
+      "step": 90
+    },
+    {
+      "epoch": 0.1834862385321101,
+      "grad_norm": 0.02230112999677658,
+      "learning_rate": 0.0009541828235992926,
+      "loss": 3.2474,
+      "step": 100
+    },
+    {
+      "epoch": 0.2018348623853211,
+      "grad_norm": 0.019073694944381714,
+      "learning_rate": 0.0009406469401200151,
+      "loss": 3.236,
+      "step": 110
+    },
+    {
+      "epoch": 0.22018348623853212,
+      "grad_norm": 0.014347660355269909,
+      "learning_rate": 0.0009254844753287492,
+      "loss": 3.2353,
+      "step": 120
+    },
+    {
+      "epoch": 0.23853211009174313,
+      "grad_norm": 0.0159588810056448,
+      "learning_rate": 0.0009087513991633924,
+      "loss": 3.2214,
+      "step": 130
+    },
+    {
+      "epoch": 0.25688073394495414,
+      "grad_norm": 0.01711818017065525,
+      "learning_rate": 0.0008905094792353917,
+      "loss": 3.2167,
+      "step": 140
+    },
+    {
+      "epoch": 0.27522935779816515,
+      "grad_norm": 0.029258986935019493,
+      "learning_rate": 0.0008708260528239789,
+      "loss": 3.2194,
+      "step": 150
+    },
+    {
+      "epoch": 0.29357798165137616,
+      "grad_norm": 0.018484625965356827,
+      "learning_rate": 0.0008497737783108237,
+      "loss": 3.21,
+      "step": 160
+    },
+    {
+      "epoch": 0.3119266055045872,
+      "grad_norm": 0.014730696566402912,
+      "learning_rate": 0.0008274303669726426,
+      "loss": 3.2112,
+      "step": 170
+    },
+    {
+      "epoch": 0.3302752293577982,
+      "grad_norm": 0.024160712957382202,
+      "learning_rate": 0.0008038782961218137,
+      "loss": 3.2021,
+      "step": 180
+    },
+    {
+      "epoch": 0.3486238532110092,
+      "grad_norm": 0.020059479400515556,
+      "learning_rate": 0.000779204504653894,
+      "loss": 3.208,
+      "step": 190
+    },
+    {
+      "epoch": 0.3669724770642202,
+      "grad_norm": 0.013730284757912159,
+      "learning_rate": 0.0007535000721258791,
+      "loss": 3.1901,
+      "step": 200
+    },
+    {
+      "epoch": 0.3853211009174312,
+      "grad_norm": 0.01747654750943184,
+      "learning_rate": 0.00072685988254984,
+      "loss": 3.193,
+      "step": 210
+    },
+    {
+      "epoch": 0.4036697247706422,
+      "grad_norm": 0.02208118513226509,
+      "learning_rate": 0.0006993822741429906,
+      "loss": 3.1997,
+      "step": 220
+    },
+    {
+      "epoch": 0.42201834862385323,
+      "grad_norm": 0.02956506237387657,
+      "learning_rate": 0.0006711686763270818,
+      "loss": 3.2017,
+      "step": 230
+    },
+    {
+      "epoch": 0.44036697247706424,
+      "grad_norm": 0.021700628101825714,
+      "learning_rate": 0.0006423232353170831,
+      "loss": 3.1906,
+      "step": 240
+    },
+    {
+      "epoch": 0.45871559633027525,
+      "grad_norm": 0.016486668959259987,
+      "learning_rate": 0.0006129524296812335,
+      "loss": 3.1824,
+      "step": 250
+    },
+    {
+      "epoch": 0.47706422018348627,
+      "grad_norm": 0.026177579537034035,
+      "learning_rate": 0.0005831646772915651,
+      "loss": 3.1964,
+      "step": 260
+    },
+    {
+      "epoch": 0.4954128440366973,
+      "grad_norm": 0.013274142518639565,
+      "learning_rate": 0.0005530699351157809,
+      "loss": 3.1816,
+      "step": 270
+    },
+    {
+      "epoch": 0.5137614678899083,
+      "grad_norm": 0.021060507744550705,
+      "learning_rate": 0.0005227792933277943,
+      "loss": 3.1824,
+      "step": 280
+    },
+    {
+      "epoch": 0.5321100917431193,
+      "grad_norm": 0.019369766116142273,
+      "learning_rate": 0.0004924045652352092,
+      "loss": 3.1697,
+      "step": 290
+    },
+    {
+      "epoch": 0.5504587155963303,
+      "grad_norm": 0.016147347167134285,
+      "learning_rate": 0.0004620578745374633,
+      "loss": 3.1699,
+      "step": 300
+    },
+    {
+      "epoch": 0.5688073394495413,
+      "grad_norm": 0.017658894881606102,
+      "learning_rate": 0.0004318512414382059,
+      "loss": 3.1789,
+      "step": 310
+    },
+    {
+      "epoch": 0.5871559633027523,
+      "grad_norm": 0.0148451654240489,
+      "learning_rate": 0.0004018961691397148,
+      "loss": 3.1797,
+      "step": 320
+    },
+    {
+      "epoch": 0.6055045871559633,
+      "grad_norm": 0.018892893567681313,
+      "learning_rate": 0.0003723032322457458,
+      "loss": 3.1651,
+      "step": 330
+    },
+    {
+      "epoch": 0.6238532110091743,
+      "grad_norm": 0.015004049055278301,
+      "learning_rate": 0.0003431816685921625,
+      "loss": 3.1706,
+      "step": 340
+    },
+    {
+      "epoch": 0.6422018348623854,
+      "grad_norm": 0.014445029199123383,
+      "learning_rate": 0.0003146389760120416,
+      "loss": 3.1801,
+      "step": 350
+    },
+    {
+      "epoch": 0.6605504587155964,
+      "grad_norm": 0.01581163890659809,
+      "learning_rate": 0.00028678051552373485,
+      "loss": 3.1777,
+      "step": 360
+    },
+    {
+      "epoch": 0.6788990825688074,
+      "grad_norm": 0.014177697710692883,
+      "learning_rate": 0.0002597091224066581,
+      "loss": 3.1673,
+      "step": 370
+    },
+    {
+      "epoch": 0.6972477064220184,
+      "grad_norm": 0.01922370120882988,
+      "learning_rate": 0.0002335247266004629,
+      "loss": 3.1813,
+      "step": 380
+    },
+    {
+      "epoch": 0.7155963302752294,
+      "grad_norm": 0.0210433192551136,
+      "learning_rate": 0.000208323983828828,
+      "loss": 3.1725,
+      "step": 390
+    },
+    {
+      "epoch": 0.7339449541284404,
+      "grad_norm": 0.01800917275249958,
+      "learning_rate": 0.0001841999188095224,
+      "loss": 3.1668,
+      "step": 400
+    },
+    {
+      "epoch": 0.7522935779816514,
+      "grad_norm": 0.01645796373486519,
+      "learning_rate": 0.00016124158186777676,
+      "loss": 3.1686,
+      "step": 410
+    },
+    {
+      "epoch": 0.7706422018348624,
+      "grad_norm": 0.0161184910684824,
+      "learning_rate": 0.0001395337202205211,
+      "loss": 3.1615,
+      "step": 420
+    },
+    {
+      "epoch": 0.7889908256880734,
+      "grad_norm": 0.016883913427591324,
+      "learning_rate": 0.00011915646514489292,
+      "loss": 3.1666,
+      "step": 430
+    },
+    {
+      "epoch": 0.8073394495412844,
+      "grad_norm": 0.016956966370344162,
+      "learning_rate": 0.00010018503618578818,
+      "loss": 3.1756,
+      "step": 440
+    },
+    {
+      "epoch": 0.8256880733944955,
+      "grad_norm": 0.015745555981993675,
+      "learning_rate": 8.268946349432583e-05,
+      "loss": 3.1753,
+      "step": 450
+    },
+    {
+      "epoch": 0.8440366972477065,
+      "grad_norm": 0.015427568927407265,
+      "learning_rate": 6.673432932217117e-05,
+      "loss": 3.165,
+      "step": 460
+    },
+    {
+      "epoch": 0.8623853211009175,
+      "grad_norm": 0.01617266982793808,
+      "learning_rate": 5.2378529625954684e-05,
+      "loss": 3.155,
+      "step": 470
+    },
+    {
+      "epoch": 0.8807339449541285,
+      "grad_norm": 0.016359878703951836,
+      "learning_rate": 3.9675056661785556e-05,
+      "loss": 3.1619,
+      "step": 480
+    },
+    {
+      "epoch": 0.8990825688073395,
+      "grad_norm": 0.011964860372245312,
+      "learning_rate": 2.8670803372379538e-05,
+      "loss": 3.1692,
+      "step": 490
+    },
+    {
+      "epoch": 0.9174311926605505,
+      "grad_norm": 0.013448962941765785,
+      "learning_rate": 1.9406390288876586e-05,
+      "loss": 3.1761,
+      "step": 500
+    },
+    {
+      "epoch": 0.9357798165137615,
+      "grad_norm": 0.014681410975754261,
+      "learning_rate": 1.191601558631461e-05,
+      "loss": 3.1631,
+      "step": 510
+    },
+    {
+      "epoch": 0.9541284403669725,
+      "grad_norm": 0.01263352669775486,
+      "learning_rate": 6.2273288462559305e-06,
+      "loss": 3.1581,
+      "step": 520
+    },
+    {
+      "epoch": 0.9724770642201835,
+      "grad_norm": 0.012895628809928894,
+      "learning_rate": 2.361328992552314e-06,
+      "loss": 3.1622,
+      "step": 530
+    },
+    {
+      "epoch": 0.9908256880733946,
+      "grad_norm": 0.011635190807282925,
+      "learning_rate": 3.322867770044202e-07,
+      "loss": 3.1642,
+      "step": 540
+    },
+    {
+      "epoch": 1.0,
+      "step": 545,
+      "total_flos": 3.409373144757043e+17,
+      "train_loss": 3.254432085019733,
+      "train_runtime": 5259.5588,
+      "train_samples_per_second": 106.117,
+      "train_steps_per_second": 0.104
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 545,
+  "num_input_tokens_seen": 29204608,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": false,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 3.409373144757043e+17,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

continuous/qwen25/pixio_vith16/pretrain/config.json ADDED Viewed

	@@ -0,0 +1,87 @@

+{
+  "add_faster_video": false,
+  "add_time_instruction": false,
+  "architectures": [
+    "Qwen2ForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "dtype": "bfloat16",
+  "eos_token_id": 151645,
+  "faster_token_stride": 10,
+  "force_sample": false,
+  "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,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "mm_hidden_size": 1280,
+  "mm_newline_position": "grid",
+  "mm_patch_merge_type": "flat",
+  "mm_projector_lr": null,
+  "mm_projector_type": "mlp2x_gelu",
+  "mm_resampler_type": null,
+  "mm_spatial_pool_mode": "bilinear",
+  "mm_spatial_pool_stride": null,
+  "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": -1,
+  "mm_vision_tower": "vtb_ssl:pixio",
+  "mm_vision_tower_lr": null,
+  "model_type": "qwen2",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 2,
+  "pad_token_id": 151643,
+  "pos_skipping_range": 4096,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "tokenizer_model_max_length": 8192,
+  "tokenizer_padding_side": "right",
+  "transformers_version": "4.57.6",
+  "use_cache": true,
+  "use_mm_proj": true,
+  "use_pos_skipping": false,
+  "use_sliding_window": false,
+  "vision_tower_pretrained": "/cache/ckpt/download/tokenizer/continuous/pixio/pixio-vith16/pixio_vith16.pth",
+  "vocab_size": 151936
+}

continuous/qwen25/pixio_vith16/pretrain/mm_projector.bin ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:4616b9e9b850693a5ff96adbbb2293c22b4bdd10a8950629c35f3c1e950aa2cc
+size 17315896

continuous/qwen25/pixio_vith16/pretrain/trainer_state.json ADDED Viewed

	@@ -0,0 +1,421 @@

+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 545,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01834862385321101,
+      "grad_norm": 0.1916215419769287,
+      "learning_rate": 0.00032142857142857147,
+      "loss": 4.8293,
+      "step": 10
+    },
+    {
+      "epoch": 0.03669724770642202,
+      "grad_norm": 0.06307043135166168,
+      "learning_rate": 0.0006785714285714287,
+      "loss": 4.2153,
+      "step": 20
+    },
+    {
+      "epoch": 0.05504587155963303,
+      "grad_norm": 0.07134393602609634,
+      "learning_rate": 0.0009999907688176173,
+      "loss": 3.9388,
+      "step": 30
+    },
+    {
+      "epoch": 0.07339449541284404,
+      "grad_norm": 0.11032756417989731,
+      "learning_rate": 0.0009988834393115766,
+      "loss": 3.7207,
+      "step": 40
+    },
+    {
+      "epoch": 0.09174311926605505,
+      "grad_norm": 0.035670001059770584,
+      "learning_rate": 0.0009959345572940772,
+      "loss": 3.3839,
+      "step": 50
+    },
+    {
+      "epoch": 0.11009174311926606,
+      "grad_norm": 0.015688125044107437,
+      "learning_rate": 0.0009911550081155983,
+      "loss": 3.3208,
+      "step": 60
+    },
+    {
+      "epoch": 0.12844036697247707,
+      "grad_norm": 0.01936701312661171,
+      "learning_rate": 0.0009845624347566433,
+      "loss": 3.2834,
+      "step": 70
+    },
+    {
+      "epoch": 0.14678899082568808,
+      "grad_norm": 0.016871247440576553,
+      "learning_rate": 0.0009761811727013547,
+      "loss": 3.2578,
+      "step": 80
+    },
+    {
+      "epoch": 0.1651376146788991,
+      "grad_norm": 0.03159899264574051,
+      "learning_rate": 0.0009660421601067666,
+      "loss": 3.2352,
+      "step": 90
+    },
+    {
+      "epoch": 0.1834862385321101,
+      "grad_norm": 0.03150615096092224,
+      "learning_rate": 0.0009541828235992926,
+      "loss": 3.2267,
+      "step": 100
+    },
+    {
+      "epoch": 0.2018348623853211,
+      "grad_norm": 0.016958773136138916,
+      "learning_rate": 0.0009406469401200151,
+      "loss": 3.2062,
+      "step": 110
+    },
+    {
+      "epoch": 0.22018348623853212,
+      "grad_norm": 0.04282128065824509,
+      "learning_rate": 0.0009254844753287492,
+      "loss": 3.2063,
+      "step": 120
+    },
+    {
+      "epoch": 0.23853211009174313,
+      "grad_norm": 0.03239241987466812,
+      "learning_rate": 0.0009087513991633924,
+      "loss": 3.1774,
+      "step": 130
+    },
+    {
+      "epoch": 0.25688073394495414,
+      "grad_norm": 0.0289929136633873,
+      "learning_rate": 0.0008905094792353917,
+      "loss": 3.1644,
+      "step": 140
+    },
+    {
+      "epoch": 0.27522935779816515,
+      "grad_norm": 0.0382227785885334,
+      "learning_rate": 0.0008708260528239789,
+      "loss": 3.1577,
+      "step": 150
+    },
+    {
+      "epoch": 0.29357798165137616,
+      "grad_norm": 0.04931199178099632,
+      "learning_rate": 0.0008497737783108237,
+      "loss": 3.1386,
+      "step": 160
+    },
+    {
+      "epoch": 0.3119266055045872,
+      "grad_norm": 0.034453608095645905,
+      "learning_rate": 0.0008274303669726426,
+      "loss": 3.1327,
+      "step": 170
+    },
+    {
+      "epoch": 0.3302752293577982,
+      "grad_norm": 0.028284989297389984,
+      "learning_rate": 0.0008038782961218137,
+      "loss": 3.1161,
+      "step": 180
+    },
+    {
+      "epoch": 0.3486238532110092,
+      "grad_norm": 0.03544061630964279,
+      "learning_rate": 0.000779204504653894,
+      "loss": 3.1122,
+      "step": 190
+    },
+    {
+      "epoch": 0.3669724770642202,
+      "grad_norm": 0.030671877786517143,
+      "learning_rate": 0.0007535000721258791,
+      "loss": 3.0882,
+      "step": 200
+    },
+    {
+      "epoch": 0.3853211009174312,
+      "grad_norm": 0.04174326732754707,
+      "learning_rate": 0.00072685988254984,
+      "loss": 3.0825,
+      "step": 210
+    },
+    {
+      "epoch": 0.4036697247706422,
+      "grad_norm": 0.035072121769189835,
+      "learning_rate": 0.0006993822741429906,
+      "loss": 3.0848,
+      "step": 220
+    },
+    {
+      "epoch": 0.42201834862385323,
+      "grad_norm": 0.02996746264398098,
+      "learning_rate": 0.0006711686763270818,
+      "loss": 3.0764,
+      "step": 230
+    },
+    {
+      "epoch": 0.44036697247706424,
+      "grad_norm": 0.027739303186535835,
+      "learning_rate": 0.0006423232353170831,
+      "loss": 3.057,
+      "step": 240
+    },
+    {
+      "epoch": 0.45871559633027525,
+      "grad_norm": 0.02627563662827015,
+      "learning_rate": 0.0006129524296812335,
+      "loss": 3.0426,
+      "step": 250
+    },
+    {
+      "epoch": 0.47706422018348627,
+      "grad_norm": 0.03686441481113434,
+      "learning_rate": 0.0005831646772915651,
+      "loss": 3.0502,
+      "step": 260
+    },
+    {
+      "epoch": 0.4954128440366973,
+      "grad_norm": 0.03722156956791878,
+      "learning_rate": 0.0005530699351157809,
+      "loss": 3.0259,
+      "step": 270
+    },
+    {
+      "epoch": 0.5137614678899083,
+      "grad_norm": 0.031570907682180405,
+      "learning_rate": 0.0005227792933277943,
+      "loss": 3.0265,
+      "step": 280
+    },
+    {
+      "epoch": 0.5321100917431193,
+      "grad_norm": 0.03186466172337532,
+      "learning_rate": 0.0004924045652352092,
+      "loss": 3.0036,
+      "step": 290
+    },
+    {
+      "epoch": 0.5504587155963303,
+      "grad_norm": 0.03241283819079399,
+      "learning_rate": 0.0004620578745374633,
+      "loss": 3.003,
+      "step": 300
+    },
+    {
+      "epoch": 0.5688073394495413,
+      "grad_norm": 0.034215763211250305,
+      "learning_rate": 0.0004318512414382059,
+      "loss": 3.0071,
+      "step": 310
+    },
+    {
+      "epoch": 0.5871559633027523,
+      "grad_norm": 0.02961408719420433,
+      "learning_rate": 0.0004018961691397148,
+      "loss": 2.9993,
+      "step": 320
+    },
+    {
+      "epoch": 0.6055045871559633,
+      "grad_norm": 0.03514479473233223,
+      "learning_rate": 0.0003723032322457458,
+      "loss": 2.9774,
+      "step": 330
+    },
+    {
+      "epoch": 0.6238532110091743,
+      "grad_norm": 0.038187913596630096,
+      "learning_rate": 0.0003431816685921625,
+      "loss": 2.9829,
+      "step": 340
+    },
+    {
+      "epoch": 0.6422018348623854,
+      "grad_norm": 0.03063701093196869,
+      "learning_rate": 0.0003146389760120416,
+      "loss": 2.9857,
+      "step": 350
+    },
+    {
+      "epoch": 0.6605504587155964,
+      "grad_norm": 0.032103411853313446,
+      "learning_rate": 0.00028678051552373485,
+      "loss": 2.9868,
+      "step": 360
+    },
+    {
+      "epoch": 0.6788990825688074,
+      "grad_norm": 0.02651306986808777,
+      "learning_rate": 0.0002597091224066581,
+      "loss": 2.9707,
+      "step": 370
+    },
+    {
+      "epoch": 0.6972477064220184,
+      "grad_norm": 0.03026103600859642,
+      "learning_rate": 0.0002335247266004629,
+      "loss": 2.9825,
+      "step": 380
+    },
+    {
+      "epoch": 0.7155963302752294,
+      "grad_norm": 0.04110582545399666,
+      "learning_rate": 0.000208323983828828,
+      "loss": 2.9713,
+      "step": 390
+    },
+    {
+      "epoch": 0.7339449541284404,
+      "grad_norm": 0.03112679533660412,
+      "learning_rate": 0.0001841999188095224,
+      "loss": 2.962,
+      "step": 400
+    },
+    {
+      "epoch": 0.7522935779816514,
+      "grad_norm": 0.029198283329606056,
+      "learning_rate": 0.00016124158186777676,
+      "loss": 2.9586,
+      "step": 410
+    },
+    {
+      "epoch": 0.7706422018348624,
+      "grad_norm": 0.025181075558066368,
+      "learning_rate": 0.0001395337202205211,
+      "loss": 2.9555,
+      "step": 420
+    },
+    {
+      "epoch": 0.7889908256880734,
+      "grad_norm": 0.025397829711437225,
+      "learning_rate": 0.00011915646514489292,
+      "loss": 2.9556,
+      "step": 430
+    },
+    {
+      "epoch": 0.8073394495412844,
+      "grad_norm": 0.03358311951160431,
+      "learning_rate": 0.00010018503618578818,
+      "loss": 2.966,
+      "step": 440
+    },
+    {
+      "epoch": 0.8256880733944955,
+      "grad_norm": 0.022470394149422646,
+      "learning_rate": 8.268946349432583e-05,
+      "loss": 2.9605,
+      "step": 450
+    },
+    {
+      "epoch": 0.8440366972477065,
+      "grad_norm": 0.022897036746144295,
+      "learning_rate": 6.673432932217117e-05,
+      "loss": 2.9493,
+      "step": 460
+    },
+    {
+      "epoch": 0.8623853211009175,
+      "grad_norm": 0.023667413741350174,
+      "learning_rate": 5.2378529625954684e-05,
+      "loss": 2.9382,
+      "step": 470
+    },
+    {
+      "epoch": 0.8807339449541285,
+      "grad_norm": 0.022257696837186813,
+      "learning_rate": 3.9675056661785556e-05,
+      "loss": 2.947,
+      "step": 480
+    },
+    {
+      "epoch": 0.8990825688073395,
+      "grad_norm": 0.02072245441377163,
+      "learning_rate": 2.8670803372379538e-05,
+      "loss": 2.9526,
+      "step": 490
+    },
+    {
+      "epoch": 0.9174311926605505,
+      "grad_norm": 0.022944586351513863,
+      "learning_rate": 1.9406390288876586e-05,
+      "loss": 2.9604,
+      "step": 500
+    },
+    {
+      "epoch": 0.9357798165137615,
+      "grad_norm": 0.019592221826314926,
+      "learning_rate": 1.191601558631461e-05,
+      "loss": 2.9464,
+      "step": 510
+    },
+    {
+      "epoch": 0.9541284403669725,
+      "grad_norm": 0.02172590233385563,
+      "learning_rate": 6.2273288462559305e-06,
+      "loss": 2.9423,
+      "step": 520
+    },
+    {
+      "epoch": 0.9724770642201835,
+      "grad_norm": 0.019899938255548477,
+      "learning_rate": 2.361328992552314e-06,
+      "loss": 2.9446,
+      "step": 530
+    },
+    {
+      "epoch": 0.9908256880733946,
+      "grad_norm": 0.01947406865656376,
+      "learning_rate": 3.322867770044202e-07,
+      "loss": 2.9479,
+      "step": 540
+    },
+    {
+      "epoch": 1.0,
+      "step": 545,
+      "total_flos": 3.409373144757043e+17,
+      "train_loss": 3.137567663630214,
+      "train_runtime": 5267.7082,
+      "train_samples_per_second": 105.953,
+      "train_steps_per_second": 0.103
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 545,
+  "num_input_tokens_seen": 29204608,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": false,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 3.409373144757043e+17,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

continuous/qwen25/pixio_vitl16/finetune/added_tokens.json ADDED Viewed

	@@ -0,0 +1,24 @@

+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

continuous/qwen25/pixio_vitl16/finetune/chat_template.jinja ADDED Viewed

	@@ -0,0 +1,54 @@

+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- messages[0]['content'] }}
+    {%- else %}
+        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
+    {%- endif %}
+    {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
+    {%- else %}
+        {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role }}
+        {%- if message.content %}
+            {{- '\n' + message.content }}
+        {%- endif %}
+        {%- for tool_call in message.tool_calls %}
+            {%- if tool_call.function is defined %}
+                {%- set tool_call = tool_call.function %}
+            {%- endif %}
+            {{- '\n<tool_call>\n{"name": "' }}
+            {{- tool_call.name }}
+            {{- '", "arguments": ' }}
+            {{- tool_call.arguments | tojson }}
+            {{- '}\n</tool_call>' }}
+        {%- endfor %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+{%- endif %}

continuous/qwen25/pixio_vitl16/finetune/config.json ADDED Viewed

	@@ -0,0 +1,87 @@

+{
+  "add_faster_video": false,
+  "add_time_instruction": false,
+  "architectures": [
+    "LlavaQwenForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "dtype": "bfloat16",
+  "eos_token_id": 151645,
+  "faster_token_stride": 10,
+  "force_sample": false,
+  "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,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
+  "max_position_embeddings": 32768,
+  "max_window_layers": 21,
+  "mm_hidden_size": 1024,
+  "mm_newline_position": "grid",
+  "mm_patch_merge_type": "flat",
+  "mm_projector_lr": 0.0001,
+  "mm_projector_type": "mlp2x_gelu",
+  "mm_resampler_type": null,
+  "mm_spatial_pool_mode": "bilinear",
+  "mm_spatial_pool_stride": null,
+  "mm_tunable_parts": "mm_mlp_adapter,mm_language_model",
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -1,
+  "mm_vision_tower": "vtb_ssl:pixio",
+  "mm_vision_tower_lr": null,
+  "model_type": "qwen2",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 28,
+  "num_key_value_heads": 2,
+  "pad_token_id": 151643,
+  "pos_skipping_range": 4096,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": true,
+  "tokenizer_model_max_length": 8192,
+  "tokenizer_padding_side": "right",
+  "transformers_version": "4.57.6",
+  "use_cache": true,
+  "use_mm_proj": true,
+  "use_pos_skipping": false,
+  "use_sliding_window": false,
+  "vision_tower_pretrained": "/cache/ckpt/download/tokenizer/continuous/pixio/pixio-vitl16/pixio_vitl16.pth",
+  "vocab_size": 151936
+}

continuous/qwen25/pixio_vitl16/finetune/generation_config.json ADDED Viewed

	@@ -0,0 +1,13 @@

+{
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.1,
+  "temperature": 0.7,
+  "top_k": 20,
+  "top_p": 0.8,
+  "transformers_version": "4.57.6"
+}

continuous/qwen25/pixio_vitl16/finetune/merges.txt ADDED Viewed

The diff for this file is too large to render. See raw diff