Upload folder using huggingface_hub

checkpoints/latest-checkpoint.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a4f68b9169d7419850d408c6eb75bb606ff66301540c32670503026965943496
+size 7125006107

config.json

@@ -0,0 +1,73 @@
+{
+  "dataset": {
+    "align_stage_components": [
+      "download/llava-laion-cc-sbu-558k/chat.json",
+      "download/llava-laion-cc-sbu-558k"
+    ],
+    "dataset_id": "llava-v15",
+    "dataset_resampled": true,
+    "dataset_root_dir": "data",
+    "finetune_stage_components": [
+      "/local/home/weizhiwang/data/MAmmoTH-VL-Instruct-12M/mammoth_si_10M_simple.jsonl",
+      "/share/edc/home/weizhiwang/data/MAmmoTH-VL-Instruct-12M/single_image_data"
+    ],
+    "max_num_images": 6,
+    "min_num_images": 1,
+    "train_num_samples": 200000,
+    "type": "llava-v15",
+    "workers": 4
+  },
+  "model": {
+    "align_epochs": 1,
+    "align_global_batch_size": 256,
+    "align_learning_rate": 0.001,
+    "align_lr_scheduler_type": "linear-warmup+cosine-decay",
+    "align_max_grad_norm": 1.0,
+    "align_max_steps": null,
+    "align_per_device_batch_size": 16,
+    "align_train_strategy": "fsdp-shard-grad-op",
+    "align_warmup_ratio": 0.03,
+    "align_weight_decay": 0.0,
+    "arch_specifier": "full-align+729-avgpool",
+    "enable_gradient_checkpointing": true,
+    "enable_mixed_precision_training": true,
+    "finetune_epochs": 1,
+    "finetune_global_batch_size": 128,
+    "finetune_learning_rate": 2e-05,
+    "finetune_lr_scheduler_type": "linear-warmup+cosine-decay",
+    "finetune_max_grad_norm": 1.0,
+    "finetune_max_steps": null,
+    "finetune_per_device_batch_size": 2,
+    "finetune_train_strategy": "fsdp-full-shard",
+    "finetune_warmup_ratio": 0.03,
+    "finetune_weight_decay": 0.1,
+    "image_resize_strategy": "resize-naive",
+    "llm_backbone_id": "qwen2.5-1.5b-instruct",
+    "llm_max_length": 4096,
+    "model_id": "qwen2.5-1.5b-instruct-continue-training-ccs-datacomp-mlm-filter-mammoth-10m",
+    "pretrain_epochs": 1,
+    "pretrain_global_batch_size": 512,
+    "pretrain_learning_rate": 5e-05,
+    "pretrain_lr_scheduler_type": "linear-warmup+cosine-decay",
+    "pretrain_max_grad_norm": 1.0,
+    "pretrain_max_steps": null,
+    "pretrain_per_device_batch_size": 16,
+    "pretrain_train_strategy": "fsdp-full-shard",
+    "pretrain_warmup_ratio": 0.03,
+    "pretrain_weight_decay": 0.01,
+    "reduce_in_full_precision": false,
+    "type": "one-stage+7b",
+    "vision_backbone_id": "siglip-vit-so400m-384px"
+  },
+  "mount_path": "Qwen",
+  "pretrained_checkpoint": "/local/home/weizhiwang/checkpoints/obelics+qwen2.5-1.5b-instruct-continue-training-ccs-datacomp-mlm-filter+stage-pretrain+x7/checkpoints/latest-checkpoint.pt",
+  "run_id": "qwen2.5-1.5b-instruct-continue-training-ccs-datacomp-mlm-filter-mammoth-10m+stage-finetune+x7",
+  "run_root_dir": "/share/edc/home/weizhiwang/checkpoints",
+  "seed": 7,
+  "stage": "finetune",
+  "trackers": [
+    "jsonl"
+  ],
+  "wandb_entity": null,
+  "wandb_project": "mmpretrain"
+}

config.yaml

@@ -0,0 +1,66 @@
+dataset:
+  align_stage_components:
+  - download/llava-laion-cc-sbu-558k/chat.json
+  - download/llava-laion-cc-sbu-558k
+  dataset_id: llava-v15
+  dataset_resampled: true
+  dataset_root_dir: data
+  finetune_stage_components:
+  - /local/home/weizhiwang/data/MAmmoTH-VL-Instruct-12M/mammoth_si_10M_simple.jsonl
+  - /share/edc/home/weizhiwang/data/MAmmoTH-VL-Instruct-12M/single_image_data
+  max_num_images: 6
+  min_num_images: 1
+  train_num_samples: 200000
+  type: llava-v15
+  workers: 4
+model:
+  align_epochs: 1
+  align_global_batch_size: 256
+  align_learning_rate: 0.001
+  align_lr_scheduler_type: linear-warmup+cosine-decay
+  align_max_grad_norm: 1.0
+  align_max_steps: null
+  align_per_device_batch_size: 16
+  align_train_strategy: fsdp-shard-grad-op
+  align_warmup_ratio: 0.03
+  align_weight_decay: 0.0
+  arch_specifier: full-align+729-avgpool
+  enable_gradient_checkpointing: true
+  enable_mixed_precision_training: true
+  finetune_epochs: 1
+  finetune_global_batch_size: 128
+  finetune_learning_rate: 2.0e-05
+  finetune_lr_scheduler_type: linear-warmup+cosine-decay
+  finetune_max_grad_norm: 1.0
+  finetune_max_steps: null
+  finetune_per_device_batch_size: 2
+  finetune_train_strategy: fsdp-full-shard
+  finetune_warmup_ratio: 0.03
+  finetune_weight_decay: 0.1
+  image_resize_strategy: resize-naive
+  llm_backbone_id: qwen2.5-1.5b-instruct
+  llm_max_length: 4096
+  model_id: qwen2.5-1.5b-instruct-continue-training-ccs-datacomp-mlm-filter-mammoth-10m
+  pretrain_epochs: 1
+  pretrain_global_batch_size: 512
+  pretrain_learning_rate: 5.0e-05
+  pretrain_lr_scheduler_type: linear-warmup+cosine-decay
+  pretrain_max_grad_norm: 1.0
+  pretrain_max_steps: null
+  pretrain_per_device_batch_size: 16
+  pretrain_train_strategy: fsdp-full-shard
+  pretrain_warmup_ratio: 0.03
+  pretrain_weight_decay: 0.01
+  reduce_in_full_precision: false
+  type: one-stage+7b
+  vision_backbone_id: siglip-vit-so400m-384px
+mount_path: Qwen
+pretrained_checkpoint: /local/home/weizhiwang/checkpoints/obelics+qwen2.5-1.5b-instruct-continue-training-ccs-datacomp-mlm-filter+stage-pretrain+x7/checkpoints/latest-checkpoint.pt
+run_id: qwen2.5-1.5b-instruct-continue-training-ccs-datacomp-mlm-filter-mammoth-10m+stage-finetune+x7
+run_root_dir: /share/edc/home/weizhiwang/checkpoints
+seed: 7
+stage: finetune
+trackers:
+- jsonl
+wandb_entity: null
+wandb_project: mmpretrain