RylanSchaeffer commited on
Commit
827b789
1 Parent(s): 689fe6c

Upload folder using huggingface_hub

Browse files
phi-instruct-3+4b+siglip/config.json CHANGED
@@ -40,7 +40,7 @@
40
  "image_resize_strategy": "letterbox",
41
  "llm_backbone_id": "phi-3-instruct-4b",
42
  "llm_max_length": 2048,
43
- "model_id": "phi-instruct-3+4b+siglipi",
44
  "reduce_in_full_precision": false,
45
  "type": "one-stage+7b",
46
  "vision_backbone_id": "siglip-vit-so400m-384px"
@@ -56,4 +56,4 @@
56
  ],
57
  "wandb_entity": "rylan",
58
  "wandb_project": "prismatic-vlm"
59
- }
 
40
  "image_resize_strategy": "letterbox",
41
  "llm_backbone_id": "phi-3-instruct-4b",
42
  "llm_max_length": 2048,
43
+ "model_id": "phi-instruct-3+4b+siglip",
44
  "reduce_in_full_precision": false,
45
  "type": "one-stage+7b",
46
  "vision_backbone_id": "siglip-vit-so400m-384px"
 
56
  ],
57
  "wandb_entity": "rylan",
58
  "wandb_project": "prismatic-vlm"
59
+ }
phi-instruct-3+4b+siglip/config.yaml CHANGED
@@ -36,7 +36,7 @@ model:
36
  image_resize_strategy: letterbox
37
  llm_backbone_id: phi-3-instruct-4b
38
  llm_max_length: 2048
39
- model_id: phi-instruct-3+4b+siglipi
40
  reduce_in_full_precision: false
41
  type: one-stage+7b
42
  vision_backbone_id: siglip-vit-so400m-384px
 
36
  image_resize_strategy: letterbox
37
  llm_backbone_id: phi-3-instruct-4b
38
  llm_max_length: 2048
39
+ model_id: phi-instruct-3+4b+siglip
40
  reduce_in_full_precision: false
41
  type: one-stage+7b
42
  vision_backbone_id: siglip-vit-so400m-384px
phi-instruct-3+4b+siglip/run-metrics.jsonl CHANGED
@@ -1 +1 @@
1
- {"hparams": {"dataset": {"align_stage_components": ["download/llava-laion-cc-sbu-558k/chat.json", "download/llava-laion-cc-sbu-558k"], "dataset_id": "llava-v15", "dataset_root_dir": "data", "finetune_stage_components": ["download/llava-v1.5-instruct/llava_v1_5_mix665k.json", "download/llava-v1.5-instruct"], "type": "llava-v15"}, "hf_token": ".hf_token", "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": "no-align+gelu-mlp", "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": 16, "finetune_train_strategy": "fsdp-full-shard", "finetune_warmup_ratio": 0.03, "finetune_weight_decay": 0.1, "image_resize_strategy": "letterbox", "llm_backbone_id": "phi-3-instruct-4b", "llm_max_length": 2048, "model_id": "phi-instruct-3+4b+siglipi", "reduce_in_full_precision": false, "type": "one-stage+7b", "vision_backbone_id": "siglip-vit-so400m-384px"}, "pretrained_checkpoint": null, "run_id": "phi-instruct-3+4b+siglipi+stage-finetune+x7", "run_root_dir": "runs", "seed": 7, "stage": "finetune", "trackers": ["jsonl", "wandb"], "wandb_entity": "rylan", "wandb_project": "prismatic-vlm"}, "run_id": "phi-instruct-3+4b+siglipi+stage-finetune+x7"}
 
1
+ {"hparams": {"dataset": {"align_stage_components": ["download/llava-laion-cc-sbu-558k/chat.json", "download/llava-laion-cc-sbu-558k"], "dataset_id": "llava-v15", "dataset_root_dir": "data", "finetune_stage_components": ["download/llava-v1.5-instruct/llava_v1_5_mix665k.json", "download/llava-v1.5-instruct"], "type": "llava-v15"}, "hf_token": ".hf_token", "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": "no-align+gelu-mlp", "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": 16, "finetune_train_strategy": "fsdp-full-shard", "finetune_warmup_ratio": 0.03, "finetune_weight_decay": 0.1, "image_resize_strategy": "letterbox", "llm_backbone_id": "phi-3-instruct-4b", "llm_max_length": 2048, "model_id": "phi-instruct-3+4b+siglip", "reduce_in_full_precision": false, "type": "one-stage+7b", "vision_backbone_id": "siglip-vit-so400m-384px"}, "pretrained_checkpoint": null, "run_id": "phi-instruct-3+4b+siglip+stage-finetune+x7", "run_root_dir": "runs", "seed": 7, "stage": "finetune", "trackers": ["jsonl", "wandb"], "wandb_entity": "rylan", "wandb_project": "prismatic-vlm"}, "run_id": "phi-instruct-3+4b+siglip+stage-finetune+x7"}