Upload folder using huggingface_hub

.gitattributes

@@ -63,3 +63,64 @@ sft/smoe_refinev3_665k_llava/logs/0418_1721_llava..._pope_llava_model_args_9259d
 sft/smoe_refinev3_665k_llava/logs/0418_1721_llava..._pope_llava_model_args_9259d6/textvqa_val.json filter=lfs diff=lfs merge=lfs -text
 sft/smoe_refinev3_665k_llava/logs/0418_1748_llava_v1.5_gqa_llava_model_args_9259d6/gqa.json filter=lfs diff=lfs merge=lfs -text
 sft_pretrain/Full_competesmoev30/checkpoint-1040/logs/0512_0233_llava...a_val_llava_model_args_f58958/textvqa_val.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/0.02_moce/trainer_state.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1103_llava...mstar_llava_model_args_16dc6f/mmbench_en_dev.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1103_llava...mstar_llava_model_args_16dc6f/submissions/mmbench_en_dev_results.xlsx filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1110_llava..._pope_llava_model_args_16dc6f/mmerealworld_lite.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1110_llava..._pope_llava_model_args_16dc6f/pope.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1110_llava..._pope_llava_model_args_16dc6f/textvqa_val.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1130_llava_v1.5_gqa_llava_model_args_16dc6f/gqa.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_norm/trainer_state.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_relu/logs/0417_0450_llava...mstar_llava_model_args_696dbd/mmbench_en_dev.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_relu/logs/0417_0450_llava...mstar_llava_model_args_696dbd/submissions/mmbench_en_dev_results.xlsx filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_relu/logs/0417_0459_llava..._pope_llava_model_args_696dbd/mmerealworld_lite.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_relu/logs/0417_0459_llava..._pope_llava_model_args_696dbd/pope.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_relu/logs/0417_0459_llava..._pope_llava_model_args_696dbd/textvqa_val.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_relu/logs/0417_0514_llava_v1.5_gqa_llava_model_args_696dbd/gqa.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_relu/trainer_state.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_sigmoid/logs/0417_0523_llava...mstar_llava_model_args_2153fb/mmbench_en_dev.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_sigmoid/logs/0417_0523_llava...mstar_llava_model_args_2153fb/submissions/mmbench_en_dev_results.xlsx filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_sigmoid/logs/0417_0532_llava..._pope_llava_model_args_2153fb/mmerealworld_lite.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_sigmoid/logs/0417_0532_llava..._pope_llava_model_args_2153fb/pope.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_sigmoid/logs/0417_0532_llava..._pope_llava_model_args_2153fb/textvqa_val.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_sigmoid/logs/0417_0547_llava_v1.5_gqa_llava_model_args_2153fb/gqa.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_sigmoid/trainer_state.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_silu/logs/0416_2257_llava...mstar_llava_model_args_a34f80/mmbench_en_dev.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_silu/logs/0416_2257_llava...mstar_llava_model_args_a34f80/submissions/mmbench_en_dev_results.xlsx filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_silu/logs/0417_0411_llava...mstar_llava_model_args_a34f80/mmbench_en_dev.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_silu/logs/0417_0411_llava...mstar_llava_model_args_a34f80/submissions/mmbench_en_dev_results.xlsx filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_silu/logs/0417_0418_llava..._pope_llava_model_args_a34f80/mmerealworld_lite.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_silu/logs/0417_0418_llava..._pope_llava_model_args_a34f80/pope.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_silu/logs/0417_0418_llava..._pope_llava_model_args_a34f80/textvqa_val.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_silu/logs/0417_0441_llava_v1.5_gqa_llava_model_args_a34f80/gqa.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/act/moce_silu/trainer_state.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/hrloss/0.03_moce/logs/gqa/0404_0654_llava_v1.5_gqa_llava_model_args_b4e935/gqa.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/hrloss/0.03_moce/trainer_state.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/remove_competition/logs/ai2d,mmbench_en_dev,mmmu_val,mmstar/0404_2022_llava...mstar_llava_model_args_c2354b/mmbench_en_dev.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/remove_competition/logs/ai2d,mmbench_en_dev,mmmu_val,mmstar/0404_2022_llava...mstar_llava_model_args_c2354b/submissions/mmbench_en_dev_results.xlsx filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/remove_competition/logs/gqa/0405_2347_llava_v1.5_gqa_llava_model_args_c2354b/gqa.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/remove_competition/logs/ocrbench,textvqa_val,mmerealworld_lite,pope/0404_2045_llava..._pope_llava_model_args_c2354b/mmerealworld_lite.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/remove_competition/logs/ocrbench,textvqa_val,mmerealworld_lite,pope/0404_2045_llava..._pope_llava_model_args_c2354b/pope.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/remove_competition/logs/ocrbench,textvqa_val,mmerealworld_lite,pope/0404_2045_llava..._pope_llava_model_args_c2354b/textvqa_val.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/remove_diversity/logs/ai2d,mmbench_en_dev,mmmu_val,mmstar/0405_2242_llava...mstar_llava_model_args_6f8127/mmbench_en_dev.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/remove_diversity/logs/ai2d,mmbench_en_dev,mmmu_val,mmstar/0405_2242_llava...mstar_llava_model_args_6f8127/submissions/mmbench_en_dev_results.xlsx filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/remove_diversity/logs/gqa/0405_2242_llava_v1.5_gqa_llava_model_args_6f8127/gqa.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/remove_diversity/logs/ocrbench,textvqa_val,mmerealworld_lite,pope/0405_2253_llava..._pope_llava_model_args_6f8127/mmerealworld_lite.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/remove_diversity/logs/ocrbench,textvqa_val,mmerealworld_lite,pope/0405_2253_llava..._pope_llava_model_args_6f8127/pope.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/remove_diversity/logs/ocrbench,textvqa_val,mmerealworld_lite,pope/0405_2253_llava..._pope_llava_model_args_6f8127/textvqa_val.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/remove_diversity/trainer_state.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/scheduler/0.03_moce/logs/gqa/0404_0741_llava_v1.5_gqa_llava_model_args_42bfb8/gqa.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/scheduler/0.03_moce/trainer_state.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/scheduler/0.05_moce/logs/gqa/0404_0709_llava_v1.5_gqa_llava_model_args_b48daf/gqa.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/scheduler/0.05_moce/trainer_state.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/scheduler/0.09_moce/logs/gqa/0404_0610_llava_v1.5_gqa_llava_model_args_84bad3/gqa.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/scheduler/0.09_moce/trainer_state.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/theta/0.2_moce/logs/gqa/0404_0504_llava_v1.5_gqa_llava_model_args_34716b/gqa.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/theta/0.2_moce/logs/mmerealworld_lite,ocrbench/0404_0452_llava...bench_llava_model_args_34716b/mmerealworld_lite.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/theta/0.3_moce/logs/ai2d,mmbench_en_dev,mmmu_val,mmstar/0404_0418_llava...mstar_llava_model_args_99e268/mmbench_en_dev.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/theta/0.3_moce/logs/ai2d,mmbench_en_dev,mmmu_val,mmstar/0404_0418_llava...mstar_llava_model_args_99e268/submissions/mmbench_en_dev_results.xlsx filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/theta/0.3_moce/logs/gqa/0404_1451_llava_v1.5_gqa_llava_model_args_99e268/gqa.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/theta/0.3_moce/logs/ocrbench,textvqa_val,mmerealworld_lite,pope/0404_0428_llava..._pope_llava_model_args_99e268/mmerealworld_lite.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/theta/0.3_moce/logs/ocrbench,textvqa_val,mmerealworld_lite,pope/0404_0428_llava..._pope_llava_model_args_99e268/pope.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/theta/0.3_moce/logs/ocrbench,textvqa_val,mmerealworld_lite,pope/0404_0428_llava..._pope_llava_model_args_99e268/textvqa_val.json filter=lfs diff=lfs merge=lfs -text
+CompeteSMoE/ablation_moce/theta/0.3_moce/trainer_state.json filter=lfs diff=lfs merge=lfs -text

CompeteSMoE/ablation_moce/0.02_moce/added_tokens.json

@@ -0,0 +1,13 @@
+{
+  "<|assistant|>": 32001,
+  "<|endoftext|>": 32000,
+  "<|end|>": 32007,
+  "<|placeholder1|>": 32002,
+  "<|placeholder2|>": 32003,
+  "<|placeholder3|>": 32004,
+  "<|placeholder4|>": 32005,
+  "<|placeholder5|>": 32008,
+  "<|placeholder6|>": 32009,
+  "<|system|>": 32006,
+  "<|user|>": 32010
+}

CompeteSMoE/ablation_moce/0.02_moce/config.json

@@ -0,0 +1,200 @@
+{
+  "_name_or_path": "/cm/archive/namnv78/checkpoints/phi35-siglip224/pft",
+  "architectures": [
+    "LlavaPhiForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "auto_map": {
+    "AutoConfig": "configuration_phi3.Phi3Config",
+    "AutoModelForCausalLM": "modeling_phi3.Phi3ForCausalLM"
+  },
+  "bal_comp_loss_coef": 0.005,
+  "balance_loss_coef": 0.01,
+  "bos_token_id": 1,
+  "clip_smoe": true,
+  "diversity_loss_coef": 0.005,
+  "dropout": false,
+  "e_loss_coef": 0.001,
+  "embd_pdrop": 0.0,
+  "entropy_advance_loss": false,
+  "eos_token_id": 32000,
+  "freeze_backbone": false,
+  "freeze_mm_mlp_adapter": false,
+  "hidden_act": "silu",
+  "hidden_size": 3072,
+  "hybrid": true,
+  "image_aspect_ratio": "pad",
+  "init_weight": true,
+  "initializer_range": 0.02,
+  "intermediate_size": 8192,
+  "is_cosine": false,
+  "is_norm_weight": false,
+  "local_rank": 0,
+  "loss1": "balanceloss",
+  "loss2": "zloss",
+  "luna": false,
+  "max_compete_in_iter": 9,
+  "max_position_embeddings": 131072,
+  "mlp_smoe": true,
+  "mm_hidden_size": 1152,
+  "mm_patch_merge_type": "flat",
+  "mm_projector_lr": null,
+  "mm_projector_type": "moe",
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -2,
+  "mm_vision_tower": "google/siglip-so400m-patch14-224",
+  "model_name_or_path": "/cm/archive/namnv78/checkpoints/phi35-siglip224/pft",
+  "model_type": "llava_phi",
+  "moe_name": "moce",
+  "norm_softmax": false,
+  "normalization": true,
+  "num_attention_heads": 32,
+  "num_experts": 4,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 32,
+  "num_layers": 3,
+  "num_selected": 2,
+  "number_of_previous_tokens": 2,
+  "original_max_position_embeddings": 4096,
+  "pad_token_id": 32000,
+  "pretrain_mm_mlp_adapter": "/cm/archive/namnv78/checkpoints/phi35-siglip224/pft/mm_projector.bin",
+  "rate_compete": 0.2,
+  "rate_flip": 0.07,
+  "resid_pdrop": 0.0,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "long_factor": [
+      1.0800000429153442,
+      1.1100000143051147,
+      1.1399999856948853,
+      1.340000033378601,
+      1.5899999141693115,
+      1.600000023841858,
+      1.6200000047683716,
+      2.620000123977661,
+      3.2300000190734863,
+      3.2300000190734863,
+      4.789999961853027,
+      7.400000095367432,
+      7.700000286102295,
+      9.09000015258789,
+      12.199999809265137,
+      17.670000076293945,
+      24.46000099182129,
+      28.57000160217285,
+      30.420001983642578,
+      30.840002059936523,
+      32.590003967285156,
+      32.93000411987305,
+      42.320003509521484,
+      44.96000289916992,
+      50.340003967285156,
+      50.45000457763672,
+      57.55000305175781,
+      57.93000411987305,
+      58.21000289916992,
+      60.1400032043457,
+      62.61000442504883,
+      62.62000274658203,
+      62.71000289916992,
+      63.1400032043457,
+      63.1400032043457,
+      63.77000427246094,
+      63.93000411987305,
+      63.96000289916992,
+      63.970001220703125,
+      64.02999877929688,
+      64.06999969482422,
+      64.08000183105469,
+      64.12000274658203,
+      64.41000366210938,
+      64.4800033569336,
+      64.51000213623047,
+      64.52999877929688,
+      64.83999633789062
+    ],
+    "short_factor": [
+      1.0,
+      1.0199999809265137,
+      1.0299999713897705,
+      1.0299999713897705,
+      1.0499999523162842,
+      1.0499999523162842,
+      1.0499999523162842,
+      1.0499999523162842,
+      1.0499999523162842,
+      1.0699999332427979,
+      1.0999999046325684,
+      1.1099998950958252,
+      1.1599998474121094,
+      1.1599998474121094,
+      1.1699998378753662,
+      1.2899998426437378,
+      1.339999794960022,
+      1.679999828338623,
+      1.7899998426437378,
+      1.8199998140335083,
+      1.8499997854232788,
+      1.8799997568130493,
+      1.9099997282028198,
+      1.9399996995925903,
+      1.9899996519088745,
+      2.0199997425079346,
+      2.0199997425079346,
+      2.0199997425079346,
+      2.0199997425079346,
+      2.0199997425079346,
+      2.0199997425079346,
+      2.0299997329711914,
+      2.0299997329711914,
+      2.0299997329711914,
+      2.0299997329711914,
+      2.0299997329711914,
+      2.0299997329711914,
+      2.0299997329711914,
+      2.0299997329711914,
+      2.0299997329711914,
+      2.0799996852874756,
+      2.0899996757507324,
+      2.189999580383301,
+      2.2199995517730713,
+      2.5899994373321533,
+      2.729999542236328,
+      2.749999523162842,
+      2.8399994373321533
+    ],
+    "type": "longrope"
+  },
+  "rope_theta": 10000.0,
+  "router_loss_coef": 0.02,
+  "router_theta": 0.1,
+  "router_z_loss_coef": 0.001,
+  "scales": [
+    1,
+    3
+  ],
+  "sliding_window": 262144,
+  "sparse_upcycling": true,
+  "strategy_train": "base",
+  "tie_word_embeddings": false,
+  "tokenizer_model_max_length": 2048,
+  "tokenizer_padding_side": "right",
+  "topk_max": 2,
+  "topk_min": 1,
+  "torch_dtype": "bfloat16",
+  "training": true,
+  "transformers_version": "4.43.0",
+  "tune_mm_mlp_adapter": false,
+  "unit_test": true,
+  "use_cache": true,
+  "use_mm_proj": true,
+  "use_old": false,
+  "version": "phi35",
+  "vision_tower": "google/siglip-so400m-patch14-224",
+  "vision_tower_dir": "/cm/archive/namnv78/checkpoints/phi35-siglip224/pft/clip.bin",
+  "vocab_size": 32064,
+  "warm_up": 0.05
+}

CompeteSMoE/ablation_moce/0.02_moce/generation_config.json

@@ -0,0 +1,12 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "do_sample": true,
+  "eos_token_id": [
+    32007,
+    32001,
+    32000
+  ],
+  "pad_token_id": 32000,
+  "transformers_version": "4.43.0"
+}

CompeteSMoE/ablation_moce/0.02_moce/model-00001-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:96311d5c497a7b3e1238fc9c78705d3fccd375284fcdfbb79c614bc44671fb21
+size 4972489328

CompeteSMoE/ablation_moce/0.02_moce/model-00002-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:06a9b0f75ba4af3b5be197bb7ee742f04c8269a7e4ce47ed82e9df74b29c8bc0
+size 4985976068

CompeteSMoE/ablation_moce/0.02_moce/model-00003-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c93e54b35e331edbaf527cedde70214ac1c1188459437d051aed3bc545d0d7f5
+size 248943552

CompeteSMoE/ablation_moce/0.02_moce/special_tokens_map.json

@@ -0,0 +1,24 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<unk>",
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

CompeteSMoE/ablation_moce/0.02_moce/tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
+size 499723

CompeteSMoE/ablation_moce/0.02_moce/tokenizer_config.json

@@ -0,0 +1,132 @@
+{
+  "add_bos_token": false,
+  "add_eos_token": false,
+  "add_prefix_space": true,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": false
+    },
+    "32000": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "32001": {
+      "content": "<|assistant|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32002": {
+      "content": "<|placeholder1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32003": {
+      "content": "<|placeholder2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32004": {
+      "content": "<|placeholder3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32005": {
+      "content": "<|placeholder4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32006": {
+      "content": "<|system|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32007": {
+      "content": "<|end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32008": {
+      "content": "<|placeholder5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32009": {
+      "content": "<|placeholder6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32010": {
+      "content": "<|user|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "chat_template": "{% for message in messages %}{% if message['role'] == 'system' and message['content'] %}{{'<|system|>\n' + message['content'] + '<|end|>\n'}}{% elif message['role'] == 'user' %}{{'<|user|>\n' + message['content'] + '<|end|>\n'}}{% elif message['role'] == 'assistant' %}{{'<|assistant|>\n' + message['content'] + '<|end|>\n'}}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ '<|assistant|>\n' }}{% else %}{{ eos_token }}{% endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "legacy": false,
+  "model_max_length": 2048,
+  "pad_token": "<unk>",
+  "padding_side": "right",
+  "sp_model_kwargs": {},
+  "spaces_between_special_tokens": false,
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": false
+}

CompeteSMoE/ablation_moce/0.02_moce/trainer_state.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:41d4f2ddf2425262649692c2378f18875c0eb7319706020d816a5cdbe2d1b9ce
+size 11498216

CompeteSMoE/ablation_moce/0.02_moce/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4f1b7cc021baddb264941beb3099b86bf3fe0472139c26df2dcfeb4998e59fcd
+size 7928

CompeteSMoE/ablation_moce/act/moce_norm/added_tokens.json

@@ -0,0 +1,13 @@
+{
+  "<|assistant|>": 32001,
+  "<|endoftext|>": 32000,
+  "<|end|>": 32007,
+  "<|placeholder1|>": 32002,
+  "<|placeholder2|>": 32003,
+  "<|placeholder3|>": 32004,
+  "<|placeholder4|>": 32005,
+  "<|placeholder5|>": 32008,
+  "<|placeholder6|>": 32009,
+  "<|system|>": 32006,
+  "<|user|>": 32010
+}

CompeteSMoE/ablation_moce/act/moce_norm/config.json

@@ -0,0 +1,200 @@
+{
+  "_name_or_path": "/cm/archive/namnv78/checkpoints/phi35-siglip224/pft",
+  "architectures": [
+    "LlavaPhiForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "auto_map": {
+    "AutoConfig": "configuration_phi3.Phi3Config",
+    "AutoModelForCausalLM": "modeling_phi3.Phi3ForCausalLM"
+  },
+  "bal_comp_loss_coef": 0.005,
+  "balance_loss_coef": 0.01,
+  "bos_token_id": 1,
+  "clip_smoe": true,
+  "diversity_loss_coef": 0.005,
+  "dropout": false,
+  "e_loss_coef": 0.001,
+  "embd_pdrop": 0.0,
+  "entropy_advance_loss": false,
+  "eos_token_id": 32000,
+  "freeze_backbone": false,
+  "freeze_mm_mlp_adapter": false,
+  "hidden_act": "silu",
+  "hidden_size": 3072,
+  "hybrid": true,
+  "image_aspect_ratio": "pad",
+  "init_weight": true,
+  "initializer_range": 0.02,
+  "intermediate_size": 8192,
+  "is_cosine": false,
+  "is_norm_weight": false,
+  "local_rank": 0,
+  "loss1": "balanceloss",
+  "loss2": "zloss",
+  "luna": false,
+  "max_compete_in_iter": 9,
+  "max_position_embeddings": 131072,
+  "mlp_smoe": true,
+  "mm_hidden_size": 1152,
+  "mm_patch_merge_type": "flat",
+  "mm_projector_lr": null,
+  "mm_projector_type": "moe",
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -2,
+  "mm_vision_tower": "google/siglip-so400m-patch14-224",
+  "model_name_or_path": "/cm/archive/namnv78/checkpoints/phi35-siglip224/pft",
+  "model_type": "llava_phi",
+  "moe_name": "moce_norm",
+  "norm_softmax": false,
+  "normalization": true,
+  "num_attention_heads": 32,
+  "num_experts": 4,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 32,
+  "num_layers": 3,
+  "num_selected": 2,
+  "number_of_previous_tokens": 2,
+  "original_max_position_embeddings": 4096,
+  "pad_token_id": 32000,
+  "pretrain_mm_mlp_adapter": "/cm/archive/namnv78/checkpoints/phi35-siglip224/pft/mm_projector.bin",
+  "rate_compete": 0.2,
+  "rate_flip": 0.07,
+  "resid_pdrop": 0.0,
+  "rms_norm_eps": 1e-05,
+  "rope_scaling": {
+    "long_factor": [
+      1.0800000429153442,
+      1.1100000143051147,
+      1.1399999856948853,
+      1.340000033378601,
+      1.5899999141693115,
+      1.600000023841858,
+      1.6200000047683716,
+      2.620000123977661,
+      3.2300000190734863,
+      3.2300000190734863,
+      4.789999961853027,
+      7.400000095367432,
+      7.700000286102295,
+      9.09000015258789,
+      12.199999809265137,
+      17.670000076293945,
+      24.46000099182129,
+      28.57000160217285,
+      30.420001983642578,
+      30.840002059936523,
+      32.590003967285156,
+      32.93000411987305,
+      42.320003509521484,
+      44.96000289916992,
+      50.340003967285156,
+      50.45000457763672,
+      57.55000305175781,
+      57.93000411987305,
+      58.21000289916992,
+      60.1400032043457,
+      62.61000442504883,
+      62.62000274658203,
+      62.71000289916992,
+      63.1400032043457,
+      63.1400032043457,
+      63.77000427246094,
+      63.93000411987305,
+      63.96000289916992,
+      63.970001220703125,
+      64.02999877929688,
+      64.06999969482422,
+      64.08000183105469,
+      64.12000274658203,
+      64.41000366210938,
+      64.4800033569336,
+      64.51000213623047,
+      64.52999877929688,
+      64.83999633789062
+    ],
+    "short_factor": [
+      1.0,
+      1.0199999809265137,
+      1.0299999713897705,
+      1.0299999713897705,
+      1.0499999523162842,
+      1.0499999523162842,
+      1.0499999523162842,
+      1.0499999523162842,
+      1.0499999523162842,
+      1.0699999332427979,
+      1.0999999046325684,
+      1.1099998950958252,
+      1.1599998474121094,
+      1.1599998474121094,
+      1.1699998378753662,
+      1.2899998426437378,
+      1.339999794960022,
+      1.679999828338623,
+      1.7899998426437378,
+      1.8199998140335083,
+      1.8499997854232788,
+      1.8799997568130493,
+      1.9099997282028198,
+      1.9399996995925903,
+      1.9899996519088745,
+      2.0199997425079346,
+      2.0199997425079346,
+      2.0199997425079346,
+      2.0199997425079346,
+      2.0199997425079346,
+      2.0199997425079346,
+      2.0299997329711914,
+      2.0299997329711914,
+      2.0299997329711914,
+      2.0299997329711914,
+      2.0299997329711914,
+      2.0299997329711914,
+      2.0299997329711914,
+      2.0299997329711914,
+      2.0299997329711914,
+      2.0799996852874756,
+      2.0899996757507324,
+      2.189999580383301,
+      2.2199995517730713,
+      2.5899994373321533,
+      2.729999542236328,
+      2.749999523162842,
+      2.8399994373321533
+    ],
+    "type": "longrope"
+  },
+  "rope_theta": 10000.0,
+  "router_loss_coef": 0.01,
+  "router_theta": 0.1,
+  "router_z_loss_coef": 0.001,
+  "scales": [
+    1,
+    3
+  ],
+  "sliding_window": 262144,
+  "sparse_upcycling": true,
+  "strategy_train": "base",
+  "tie_word_embeddings": false,
+  "tokenizer_model_max_length": 2048,
+  "tokenizer_padding_side": "right",
+  "topk_max": 2,
+  "topk_min": 1,
+  "torch_dtype": "bfloat16",
+  "training": true,
+  "transformers_version": "4.43.0",
+  "tune_mm_mlp_adapter": false,
+  "unit_test": true,
+  "use_cache": true,
+  "use_mm_proj": true,
+  "use_old": false,
+  "version": "phi35",
+  "vision_tower": "google/siglip-so400m-patch14-224",
+  "vision_tower_dir": "/cm/archive/namnv78/checkpoints/phi35-siglip224/pft/clip.bin",
+  "vocab_size": 32064,
+  "warm_up": 0.05
+}

CompeteSMoE/ablation_moce/act/moce_norm/generation_config.json

@@ -0,0 +1,12 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "do_sample": true,
+  "eos_token_id": [
+    32007,
+    32001,
+    32000
+  ],
+  "pad_token_id": 32000,
+  "transformers_version": "4.43.0"
+}

CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1103_llava...mstar_llava_model_args_16dc6f/mmbench_en_dev.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8c59b618ea6d2e7960219d7394214992a05ffd93ab2c9bf8815a83b859eb8fa2
+size 14568153

CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1103_llava...mstar_llava_model_args_16dc6f/rank0_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 0 eval done

CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1103_llava...mstar_llava_model_args_16dc6f/rank1_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 1 eval done

CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1103_llava...mstar_llava_model_args_16dc6f/rank2_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 2 eval done

CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1103_llava...mstar_llava_model_args_16dc6f/rank3_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 3 eval done

CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1103_llava...mstar_llava_model_args_16dc6f/results.json

@@ -0,0 +1,298 @@
+{
+    "results": {
+        "ai2d": {
+            "exact_match,flexible-extract": 0.6447538860103627,
+            "exact_match_stderr,flexible-extract": 0.008613771311019505,
+            "alias": "ai2d"
+        },
+        "mmbench_en_dev": {
+            "gpt_eval_score,none": 71.56357388316151,
+            "gpt_eval_score_stderr,none": "N/A",
+            "submission,none": null,
+            "submission_stderr,none": "N/A",
+            "alias": "mmbench_en_dev"
+        },
+        "mmmu_val": {
+            "mmmu_acc,none": 0.42778,
+            "mmmu_acc_stderr,none": "N/A",
+            "alias": "mmmu_val"
+        },
+        "mmstar": {
+            "coarse perception,none": 0.7293114877191748,
+            "coarse perception_stderr,none": "N/A",
+            "fine-grained perception,none": 0.39408621960206336,
+            "fine-grained perception_stderr,none": "N/A",
+            "instance reasoning,none": 0.5199079108104118,
+            "instance reasoning_stderr,none": "N/A",
+            "logical reasoning,none": 0.3621608314677622,
+            "logical reasoning_stderr,none": "N/A",
+            "math,none": 0.29857212866434996,
+            "math_stderr,none": "N/A",
+            "science & technology,none": 0.2526168422637996,
+            "science & technology_stderr,none": "N/A",
+            "alias": "mmstar"
+        }
+    },
+    "configs": {
+        "ai2d": {
+            "task": "ai2d",
+            "dataset_path": "lmms-lab/ai2d",
+            "dataset_kwargs": {
+                "token": true
+            },
+            "test_split": "test",
+            "doc_to_visual": "<function ai2d_doc_to_visual at 0x7ff7a9382550>",
+            "doc_to_text": "<function ai2d_doc_to_text at 0x7ff7a9382820>",
+            "doc_to_target": "<function ai2d_doc_to_target at 0x7ff7a9382ca0>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "exact_match",
+                    "aggregation": "mean",
+                    "higher_is_better": true,
+                    "ignore_case": true,
+                    "ignore_punctuation": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "max_new_tokens": 512,
+                "temperature": 0.0,
+                "do_sample": false,
+                "until": [
+                    "\n\n"
+                ]
+            },
+            "repeats": 1,
+            "filter_list": [
+                {
+                    "name": "flexible-extract",
+                    "filter": [
+                        {
+                            "function": "<class 'utils.MultiChoiceRegexFilter'>",
+                            "group_select": 0,
+                            "ignore_case": true,
+                            "ignore_punctuation": true,
+                            "regex_pattern": "([A-Z])\\."
+                        }
+                    ]
+                }
+            ],
+            "should_decontaminate": false,
+            "metadata": [
+                {
+                    "version": 0.0
+                }
+            ],
+            "model_specific_prompt_kwargs": {
+                "default": {
+                    "prompt_format": "mcq",
+                    "pre_prompt": "",
+                    "post_prompt": "\nAnswer with the option's letter from the given choices directly."
+                },
+                "gpt4v": {
+                    "prompt_format": "mcq",
+                    "pre_prompt": "",
+                    "post_prompt": "\nAbove choices are given in {option}. {content} format.\nPlease answer with the option letter from the given choices directly."
+                },
+                "qwen_vl": {
+                    "prompt_format": "qa",
+                    "pre_prompt": "",
+                    "post_prompt": " Answer:"
+                },
+                "xcomposer2_4khd": {
+                    "prompt_format": "mcq_xcomposer",
+                    "pre_prompt": "[UNUSED_TOKEN_146]user\nQuestion: ",
+                    "post_prompt": "[UNUSED_TOKEN_145]\n[UNUSED_TOKEN_146]assistant\nThe answer is"
+                }
+            },
+            "model_specific_target_kwargs": {
+                "default": "mcq",
+                "qwen_vl": "qa"
+            }
+        },
+        "mmbench_en_dev": {
+            "task": "mmbench_en_dev",
+            "dataset_path": "lmms-lab/MMBench",
+            "dataset_name": "en",
+            "dataset_kwargs": {
+                "token": true
+            },
+            "test_split": "dev",
+            "doc_to_visual": "<function mmbench_doc_to_visual at 0x7ff8085c3670>",
+            "doc_to_text": "<function mmbench_doc_to_text at 0x7ff8085c3c10>",
+            "doc_to_target": "answer",
+            "process_results": "<function mmbench_process_results at 0x7ff8085ce1f0>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "gpt_eval_score",
+                    "aggregation": "<function mmbench_aggregate_dev_results_eval at 0x7ff8085bddc0>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "submission",
+                    "aggregation": "<function mmbench_aggregate_dev_results_submission at 0x7ff8085c33a0>",
+                    "higher_is_better": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "until": [
+                    "ASSISTANT:"
+                ],
+                "max_new_tokens": 1024,
+                "temperature": 0.0,
+                "top_p": 1.0,
+                "num_beams": 1,
+                "do_sample": false,
+                "image_aspect_ratio": "original"
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "model_specific_prompt_kwargs": {
+                "default": {
+                    "pre_prompt": "",
+                    "post_prompt": "\nAnswer with the option's letter from the given choices directly."
+                }
+            },
+            "model_specific_generation_kwargs": {
+                "llava": {
+                    "image_aspect_ratio": "original"
+                }
+            }
+        },
+        "mmmu_val": {
+            "task": "mmmu_val",
+            "dataset_path": "lmms-lab/MMMU",
+            "test_split": "validation",
+            "doc_to_visual": "<function mmmu_doc_to_visual at 0x7ff7bd633670>",
+            "doc_to_text": "<function mmmu_doc_to_text at 0x7ff7bd442310>",
+            "doc_to_target": "answer",
+            "process_results": "<function mmmu_process_results at 0x7ff7bd35a160>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "mmmu_acc",
+                    "aggregation": "<function mmmu_aggregate_results at 0x7ff7bd2af040>",
+                    "higher_is_better": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "max_new_tokens": 128,
+                "until": [
+                    "\n\n"
+                ],
+                "image_aspect_ratio": "original"
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "metadata": [
+                {
+                    "version": 0.0
+                }
+            ],
+            "model_specific_generation_kwargs": {
+                "llava": {
+                    "image_aspect_ratio": "original"
+                }
+            }
+        },
+        "mmstar": {
+            "task": "mmstar",
+            "dataset_path": "Lin-Chen/MMStar",
+            "dataset_kwargs": {
+                "token": true
+            },
+            "test_split": "val",
+            "doc_to_visual": "<function mmstar_doc_to_visual at 0x7ff7a92dd700>",
+            "doc_to_text": "<function mmstar_doc_to_text at 0x7ff7a92ddaf0>",
+            "doc_to_target": "answer",
+            "process_results": "<function mmstar_process_results at 0x7ff7a92ddf70>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "coarse perception",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7ff7a92693a0>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "fine-grained perception",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7ff7a9269700>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "instance reasoning",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7ff7a9269a60>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "logical reasoning",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7ff7a9269dc0>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "science & technology",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7ff7a9271160>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "math",
+                    "aggregation": "<function mmstar_aggregate_results at 0x7ff7a92714c0>",
+                    "higher_is_better": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "until": [
+                    "\n\n"
+                ],
+                "do_sample": false
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "metadata": [
+                {
+                    "version": 0.0
+                }
+            ],
+            "model_specific_prompt_kwargs": {
+                "default": {
+                    "pre_prompt": "",
+                    "post_prompt": "\nAnswer with the option's letter from the given choices directly"
+                }
+            }
+        }
+    },
+    "versions": {
+        "ai2d": "Yaml",
+        "mmbench_en_dev": "Yaml",
+        "mmmu_val": "Yaml",
+        "mmstar": "Yaml"
+    },
+    "n-shot": {
+        "ai2d": 0,
+        "mmbench_en_dev": 0,
+        "mmmu_val": 0,
+        "mmstar": 0
+    },
+    "model_configs": {
+        "model": "llava",
+        "model_args": "pretrained=/cm/archive/namnv78/checkpoints/Xphi35-siglip224/ablation_moce/act/moce_norm,conv_template=phi35",
+        "batch_size": "1",
+        "device": null,
+        "limit": null,
+        "bootstrap_iters": 100000,
+        "gen_kwargs": ""
+    },
+    "git_hash": "0f29acb"
+}

CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1103_llava...mstar_llava_model_args_16dc6f/submissions/mmbench_en_dev_results.json

@@ -0,0 +1 @@
+{"overall_acc": 0.7156357388316151, "category_acc": {"action_recognition": 0.8888888888888888, "attribute_comparison": 0.7045454545454546, "attribute_recognition": 0.8918918918918919, "celebrity_recognition": 0.8383838383838383, "function_reasoning": 0.8481012658227848, "future_prediction": 0.5, "identity_reasoning": 0.9777777777777777, "image_emotion": 0.82, "image_quality": 0.5471698113207547, "image_scene": 0.9519230769230769, "image_style": 0.8301886792452831, "image_topic": 0.8333333333333334, "nature_relation": 0.5625, "object_localization": 0.5061728395061729, "ocr": 0.7435897435897436, "physical_property_reasoning": 0.5466666666666666, "physical_relation": 0.4583333333333333, "social_relation": 0.7906976744186046, "spatial_relationship": 0.28888888888888886, "structuralized_imagetext_understanding": 0.44871794871794873}, "l2_category_acc": {"attribute_reasoning": 0.7638190954773869, "coarse_perception": 0.8209459459459459, "finegrained_perception (cross-instance)": 0.6433566433566433, "finegrained_perception (instance-level)": 0.7474402730375427, "logic_reasoning": 0.4661016949152542, "relation_reasoning": 0.6260869565217392}}

CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1103_llava...mstar_llava_model_args_16dc6f/submissions/mmbench_en_dev_results.xlsx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fa8b7abf89c725d213146bc554dc7ff8e928c8dd471cf3978dfa832c386ba1ad
+size 865144

CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1110_llava..._pope_llava_model_args_16dc6f/mmerealworld_lite.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8850d3fd9a304bae0fd09ea1d117067140065d9e62de4bcf0092b0cd0eea8e7e
+size 1994104194

CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1110_llava..._pope_llava_model_args_16dc6f/pope.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b959fd0a6d8f742a8f007408650737d883365a9ef68dcc9862c7a96bc3dfe884
+size 17577612

CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1110_llava..._pope_llava_model_args_16dc6f/rank0_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 0 eval done

CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1110_llava..._pope_llava_model_args_16dc6f/rank1_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 1 eval done

CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1110_llava..._pope_llava_model_args_16dc6f/rank2_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 2 eval done

CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1110_llava..._pope_llava_model_args_16dc6f/rank3_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 3 eval done

CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1110_llava..._pope_llava_model_args_16dc6f/results.json

@@ -0,0 +1,257 @@
+{
+    "results": {
+        "mmerealworld_lite": {
+            "mme_realworld_score,none": 0.31214174048983845,
+            "mme_realworld_score_stderr,none": "N/A",
+            "alias": "mmerealworld_lite"
+        },
+        "ocrbench": {
+            "ocrbench_accuracy,none": 0.323,
+            "ocrbench_accuracy_stderr,none": "N/A",
+            "alias": "ocrbench"
+        },
+        "pope": {
+            "pope_accuracy,none": 0.8684444444444445,
+            "pope_accuracy_stderr,none": "N/A",
+            "pope_precision,none": 0.9288670460424211,
+            "pope_precision_stderr,none": "N/A",
+            "pope_recall,none": 0.798,
+            "pope_recall_stderr,none": "N/A",
+            "pope_f1_score,none": 0.858474778866842,
+            "pope_f1_score_stderr,none": "N/A",
+            "pope_yes_ratio,none": 0.5,
+            "pope_yes_ratio_stderr,none": "N/A",
+            "alias": "pope"
+        },
+        "textvqa_val": {
+            "exact_match,none": 0.4161000002264976,
+            "exact_match_stderr,none": 0.0067540270401247006,
+            "submission,none": null,
+            "submission_stderr,none": "N/A",
+            "alias": "textvqa_val"
+        }
+    },
+    "configs": {
+        "mmerealworld_lite": {
+            "task": "mmerealworld_lite",
+            "dataset_path": "yifanzhang114/MME-RealWorld-lite-lmms-eval",
+            "dataset_kwargs": {
+                "token": true
+            },
+            "test_split": "train",
+            "doc_to_visual": "<function mme_realworld_doc_to_visual at 0x7f4612ed0550>",
+            "doc_to_text": "<function mme_realworld_doc_to_text at 0x7f4612ed5160>",
+            "doc_to_target": "answer",
+            "process_results": "<function mme_realworld_process_results at 0x7f4612ed5ee0>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "mme_realworld_score",
+                    "aggregation": "<function mme_realworld_aggregate_results at 0x7f4612ed8b80>",
+                    "higher_is_better": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "max_new_tokens": 16,
+                "temperature": 0.0,
+                "top_p": 1.0,
+                "num_beams": 1,
+                "do_sample": false,
+                "until": [
+                    "\n\n"
+                ]
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "metadata": [
+                {
+                    "version": 0.0
+                }
+            ],
+            "model_specific_prompt_kwargs": {
+                "default": {
+                    "pre_prompt": "",
+                    "post_prompt": "\nSelect the best answer to the above multiple-choice question based on the image. Respond with only the letter (A, B, C, D, or E) of the correct option."
+                },
+                "gpt4v": {
+                    "pre_prompt": "",
+                    "post_prompt": "\nSelect the best answer to the above multiple-choice question based on the image. Respond with only the letter (A, B, C, D, or E) of the correct option."
+                },
+                "xcomposer2_4khd": {
+                    "pre_prompt": "[UNUSED_TOKEN_146]user\n",
+                    "post_prompt": " Answer this question with A, B, C, or D.[UNUSED_TOKEN_145]\n[UNUSED_TOKEN_146]assistant\n"
+                }
+            }
+        },
+        "ocrbench": {
+            "task": "ocrbench",
+            "dataset_path": "echo840/OCRBench",
+            "dataset_kwargs": {
+                "token": true
+            },
+            "test_split": "test",
+            "doc_to_visual": "<function ocrbench_doc_to_visual at 0x7f46278adb80>",
+            "doc_to_text": "<function ocrbench_doc_to_text at 0x7f462728d1f0>",
+            "doc_to_target": "answer",
+            "process_results": "<function ocrbench_process_results at 0x7f462728d550>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "ocrbench_accuracy",
+                    "aggregation": "<function ocrbench_aggregate_accuracy at 0x7f462728d8b0>",
+                    "higher_is_better": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "max_new_tokens": 128,
+                "temperature": 0.0,
+                "top_p": 1.0,
+                "num_beams": 1,
+                "do_sample": false,
+                "until": [
+                    "\n\n"
+                ]
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "metadata": [
+                {
+                    "version": 0.0
+                }
+            ]
+        },
+        "pope": {
+            "task": "pope",
+            "dataset_path": "lmms-lab/POPE",
+            "dataset_kwargs": {
+                "token": true
+            },
+            "test_split": "test",
+            "doc_to_visual": "<function pope_doc_to_visual at 0x7f4612f658b0>",
+            "doc_to_text": "<function pope_doc_to_text at 0x7f4612f65dc0>",
+            "doc_to_target": "answer",
+            "process_results": "<function pope_process_results at 0x7f4612f69310>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "pope_accuracy",
+                    "aggregation": "<function pope_aggregate_accuracy at 0x7f4612f69820>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "pope_precision",
+                    "aggregation": "<function pope_aggregate_precision at 0x7f4612f69d30>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "pope_recall",
+                    "aggregation": "<function pope_aggregate_recall at 0x7f4612ef0280>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "pope_f1_score",
+                    "aggregation": "<function pope_aggregate_f1_score at 0x7f4612ef0790>",
+                    "higher_is_better": true
+                },
+                {
+                    "metric": "pope_yes_ratio",
+                    "aggregation": "<function pope_aggregate_yes_ratio at 0x7f4612ef0ca0>",
+                    "higher_is_better": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "max_new_tokens": 128,
+                "temperature": 0.0,
+                "top_p": 1.0,
+                "num_beams": 1,
+                "do_sample": false,
+                "until": [
+                    "\n\n"
+                ]
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "metadata": [
+                {
+                    "version": 0.0
+                }
+            ]
+        },
+        "textvqa_val": {
+            "task": "textvqa_val",
+            "dataset_path": "lmms-lab/textvqa",
+            "test_split": "validation",
+            "doc_to_visual": "<function textvqa_doc_to_visual at 0x7f463f9a1ca0>",
+            "doc_to_text": "<function textvqa_doc_to_text at 0x7f463f9cc160>",
+            "doc_to_target": "answer",
+            "process_results": "<function textvqa_process_results at 0x7f463f9cc430>",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "exact_match",
+                    "aggregation": "mean",
+                    "higher_is_better": true,
+                    "ignore_case": true,
+                    "ignore_punctuation": true
+                },
+                {
+                    "metric": "submission",
+                    "aggregation": "<function textvqa_aggregate_submissions at 0x7f463f9a1af0>",
+                    "higher_is_better": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "until": [
+                    "ASSISTANT:"
+                ]
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "model_specific_prompt_kwargs": {
+                "default": {
+                    "pre_prompt": "",
+                    "post_prompt": "\nAnswer the question using a single word or phrase.",
+                    "ocr": false
+                },
+                "qwen_vl": {
+                    "pre_prompt": "",
+                    "post_prompt": " Answer:"
+                }
+            }
+        }
+    },
+    "versions": {
+        "mmerealworld_lite": "Yaml",
+        "ocrbench": "Yaml",
+        "pope": "Yaml",
+        "textvqa_val": "Yaml"
+    },
+    "n-shot": {
+        "mmerealworld_lite": 0,
+        "ocrbench": 0,
+        "pope": 0,
+        "textvqa_val": 0
+    },
+    "model_configs": {
+        "model": "llava",
+        "model_args": "pretrained=/cm/archive/namnv78/checkpoints/Xphi35-siglip224/ablation_moce/act/moce_norm,conv_template=phi35",
+        "batch_size": "1",
+        "device": null,
+        "limit": null,
+        "bootstrap_iters": 100000,
+        "gen_kwargs": ""
+    },
+    "git_hash": "0f29acb"
+}

CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1110_llava..._pope_llava_model_args_16dc6f/results/ocrbench_results.txt

@@ -0,0 +1,18 @@
+######################### OCRBench #############################
+Text Recognition(Total 300): 183
+---------------- Details of Recognition Score ------------------
+Regular Text Recognition(Total 50): 45
+Irregular Text Recognition(Total 50): 42
+Artistic Text Recognition(Total 50): 46
+Handwriting Recognition(Total 50): 23
+Digit String Recognition(Total 50): 8
+Non-Semantic Text Recognition(Total 50): 19
+----------------------------------------------------------------
+Scene Text-centric VQA(Total 200): 111
+----------------------------------------------------------------
+Doc-oriented VQA(Total 200): 20
+----------------------------------------------------------------
+Key Information Extraction(Total 200): 8
+Handwritten Mathematical Expression Recognition(Total 100): 1
+--------------------- Final Score ------------------------------
+Final Score(Total 1000): 323

CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1110_llava..._pope_llava_model_args_16dc6f/textvqa_val.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f84d8b696f109af83aa1087cda1c820865d1d3d3b2dcfcfcc8075a5ede131b66
+size 13139641

CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1130_llava_v1.5_gqa_llava_model_args_16dc6f/gqa.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:edf3798f6963787a3eb1088e2a45239ec1e05d8f969f17019d4e5ea2e2621ecc
+size 38369361

CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1130_llava_v1.5_gqa_llava_model_args_16dc6f/rank0_metric_eval_done.txt

@@ -0,0 +1 @@
+rank 0 eval done

CompeteSMoE/ablation_moce/act/moce_norm/logs/0418_1130_llava_v1.5_gqa_llava_model_args_16dc6f/results.json

@@ -0,0 +1,79 @@
+{
+    "results": {
+        "gqa": {
+            "exact_match,none": 0.6123390046112259,
+            "exact_match_stderr,none": 0.004344436809906456,
+            "alias": "gqa"
+        }
+    },
+    "configs": {
+        "gqa": {
+            "task": "gqa",
+            "dataset_path": "lmms-lab/GQA",
+            "dataset_name": "testdev_balanced_instructions",
+            "dataset_kwargs": {
+                "token": true
+            },
+            "test_split": "testdev",
+            "doc_to_visual": "<function gqa_doc_to_visual at 0x7f495592b1f0>",
+            "doc_to_text": "<function gqa_doc_to_text at 0x7f4955879310>",
+            "doc_to_target": "answer",
+            "description": "",
+            "target_delimiter": " ",
+            "fewshot_delimiter": "\n\n",
+            "metric_list": [
+                {
+                    "metric": "exact_match",
+                    "aggregation": "mean",
+                    "higher_is_better": true,
+                    "ignore_case": true,
+                    "ignore_punctuation": true
+                }
+            ],
+            "output_type": "generate_until",
+            "generation_kwargs": {
+                "max_new_tokens": 16,
+                "temperature": 0.0,
+                "top_p": 1.0,
+                "num_beams": 1,
+                "do_sample": false,
+                "until": [
+                    "\n\n"
+                ]
+            },
+            "repeats": 1,
+            "should_decontaminate": false,
+            "metadata": [
+                {
+                    "version": 0.0
+                }
+            ],
+            "model_specific_prompt_kwargs": {
+                "default": {
+                    "pre_prompt": "",
+                    "post_prompt": "\nAnswer the question using a single word or phrase."
+                },
+                "qwen_vl": {
+                    "pre_prompt": "",
+                    "post_prompt": " Answer:"
+                }
+            }
+        }
+    },
+    "versions": {
+        "gqa": "Yaml"
+    },
+    "n-shot": {
+        "gqa": 0
+    },
+    "model_configs": {
+        "model": "llava",
+        "model_args": "pretrained=/cm/archive/namnv78/checkpoints/Xphi35-siglip224/ablation_moce/act/moce_norm,conv_template=phi35",
+        "batch_size": "1",
+        "device": null,
+        "limit": null,
+        "bootstrap_iters": 100000,
+        "gen_kwargs": ""
+    },
+    "git_hash": "0f29acb"
+}

CompeteSMoE/ablation_moce/act/moce_norm/model-00001-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6ba472fc18f351d847aee42290d72aaabc053cb880a3f8b7cc9a10ce08908188
+size 4972489328

CompeteSMoE/ablation_moce/act/moce_norm/model-00002-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9227c414d619866fa69e10d6a2e6b16bf6e0c68d3e2afefeb3cc35be7284dd22
+size 4985976068

CompeteSMoE/ablation_moce/act/moce_norm/model-00003-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6c9c9b731bdfa931f0f7ef8c0e1b373e66fe5c6af3c79e3613fa70ea6553806e
+size 248943552

CompeteSMoE/ablation_moce/act/moce_norm/special_tokens_map.json

@@ -0,0 +1,24 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<unk>",
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

CompeteSMoE/ablation_moce/act/moce_norm/tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
+size 499723

CompeteSMoE/ablation_moce/act/moce_norm/tokenizer_config.json

@@ -0,0 +1,132 @@
+{
+  "add_bos_token": false,
+  "add_eos_token": false,
+  "add_prefix_space": true,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": false
+    },
+    "32000": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "32001": {
+      "content": "<|assistant|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32002": {
+      "content": "<|placeholder1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32003": {
+      "content": "<|placeholder2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32004": {
+      "content": "<|placeholder3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32005": {
+      "content": "<|placeholder4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32006": {
+      "content": "<|system|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32007": {
+      "content": "<|end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32008": {
+      "content": "<|placeholder5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32009": {
+      "content": "<|placeholder6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32010": {
+      "content": "<|user|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "chat_template": "{% for message in messages %}{% if message['role'] == 'system' and message['content'] %}{{'<|system|>\n' + message['content'] + '<|end|>\n'}}{% elif message['role'] == 'user' %}{{'<|user|>\n' + message['content'] + '<|end|>\n'}}{% elif message['role'] == 'assistant' %}{{'<|assistant|>\n' + message['content'] + '<|end|>\n'}}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ '<|assistant|>\n' }}{% else %}{{ eos_token }}{% endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "legacy": false,
+  "model_max_length": 2048,
+  "pad_token": "<unk>",
+  "padding_side": "right",
+  "sp_model_kwargs": {},
+  "spaces_between_special_tokens": false,
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": false
+}

CompeteSMoE/ablation_moce/act/moce_norm/trainer_state.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6ff63ca98f8ff11b11c0f6179740c751c72292f45d67c4ae1be6db01c15a7fe5
+size 11476366

CompeteSMoE/ablation_moce/act/moce_norm/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77050365b87f7736a333d4866af630145f5d14aca6f64ec5dffcccb8feda8a73
+size 7928