Upload folder using huggingface_hub

fastumi_pickandplace_qwenPI_329v4/checkpoints/steps_10000_pytorch_model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9de2530285307659ae80a4a957a109ea970b9f60060b8e41596a45bcef16ad80
+size 12444522384

fastumi_pickandplace_qwenPI_329v4/checkpoints/steps_5000_pytorch_model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0765328b7819247afcd352b221a31844f9fa33c6ec6ce2c1fdea39f2255f0f3d
+size 12444521025

fastumi_pickandplace_qwenPI_329v4/config.yaml

@@ -0,0 +1,70 @@
+datasets:
+  vla_data:
+    CoT_prompt: Your task is {instruction}. To identify the key objects for your task.
+      Locate their bounding boxes in [x1,y1,x2,y2] format.
+    data_mix: dynamic-329-v4
+    data_root_dir: playground/Datasets/FastUMI
+    dataset_py: lerobot_datasets
+    per_device_batch_size: 8
+    video_backend: torchvision_av
+framework:
+  action_model:
+    action_dim: 10
+    add_pos_embed: true
+    diffusion_model_cfg:
+      attention_head_dim: 64
+      cross_attention_dim: 2048
+      dropout: 0.2
+      final_dropout: true
+      input_embedding_dim: 2048
+      interleave_self_attention: true
+      norm_type: ada_norm
+      num_attention_heads: 32
+      num_layers: 36
+      output_dim: 1024
+      positional_embeddings: null
+    future_action_window_size: 15
+    max_seq_len: 1024
+    noise_beta_alpha: 1.5
+    noise_beta_beta: 1.0
+    noise_s: 0.999
+    num_inference_timesteps: 4
+    num_target_vision_tokens: 32
+    num_timestep_buckets: 1000
+    past_action_window_size: 0
+    state_dim: 10
+  name: QwenPI
+  qwenvl:
+    attn_implementation: flash_attention_2
+    base_vlm: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action
+    num_vl_layers: 36
+    vl_hidden_dim: 2048
+output_dir: ./results/Checkpoints/fastumi_pickandplace_qwenPI_329v4
+run_id: fastumi_pickandplace_qwenPI_329v4
+run_root_dir: ./results/Checkpoints
+seed: 42
+trainer:
+  eval_interval: 100
+  freeze_modules: null
+  gradient_accumulation_steps: 1
+  gradient_clipping: 1.0
+  is_resume: true
+  learning_rate:
+    action_model: 0.0001
+    base: 2.5e-05
+    qwen_vl_interface: 1.0e-05
+  logging_frequency: 50
+  lr_scheduler_type: cosine_with_min_lr
+  max_train_steps: 10000
+  num_warmup_steps: 5000
+  optimizer:
+    betas:
+    - 0.9
+    - 0.95
+    eps: 1.0e-08
+    weight_decay: 1.0e-08
+  save_interval: 5000
+  scheduler_specific_kwargs:
+    min_lr: 1.0e-06
+wandb_entity: 2200011093-peking-university
+wandb_project: starVLA_FastUMI_dynamic_1

fastumi_pickandplace_qwenPI_329v4/dataset_statistics.json

@@ -0,0 +1,178 @@
+{
+  "new_embodiment": {
+    "action": {
+      "mean": [
+        8.114812226267532e-05,
+        0.0011773157166317105,
+        -0.0008882488473318517,
+        0.9999980926513672,
+        0.0004620188265107572,
+        -5.512892585102236e-06,
+        -0.00045027409214526415,
+        0.9999972581863403,
+        0.0008542306604795158,
+        0.5120490193367004
+      ],
+      "std": [
+        0.002003167988732457,
+        0.0030834961216896772,
+        0.00332102132961154,
+        5.601261758867448e-05,
+        0.00517509737983346,
+        0.006088857538998127,
+        0.00517628388479352,
+        7.67255479378806e-05,
+        0.007195365149527788,
+        0.4998091757297516
+      ],
+      "max": [
+        0.008002996444702148,
+        0.01195499300956726,
+        0.009996004402637482,
+        1.0000001192092896,
+        0.03673427179455757,
+        0.05374615639448166,
+        0.02689175307750702,
+        1.0000001192092896,
+        0.06838423758745193,
+        1.0
+      ],
+      "min": [
+        -0.00818699598312378,
+        -0.008041977882385254,
+        -0.010141998529434204,
+        0.9985396265983582,
+        -0.026854006573557854,
+        -0.04442450404167175,
+        -0.03635682165622711,
+        0.997472882270813,
+        -0.06104228273034096,
+        0.0
+      ],
+      "q01": [
+        -0.00486600399017334,
+        -0.00501847505569458,
+        -0.007628242671489716,
+        0.9997773110866547,
+        -0.012261943705379964,
+        -0.015310721378773451,
+        -0.013358078151941299,
+        0.9997043466567993,
+        -0.01697732284665108,
+        0.0
+      ],
+      "q99": [
+        0.005302621722221377,
+        0.009075625538825991,
+        0.007293619811534884,
+        0.9999998211860657,
+        0.013378759231418375,
+        0.016290122829377692,
+        0.01225539410486818,
+        0.9999997615814209,
+        0.020962502509355545,
+        1.0
+      ],
+      "mask": [
+        true,
+        true,
+        true,
+        true,
+        true,
+        true,
+        true,
+        true,
+        true,
+        false
+      ],
+      "norm_modes": [
+        "min_max",
+        "min_max",
+        "min_max",
+        "none",
+        "none",
+        "none",
+        "none",
+        "none",
+        "none",
+        "binary"
+      ]
+    },
+    "state": {
+      "mean": [
+        0.27271032333374023,
+        -0.446323037147522,
+        0.1041000559926033,
+        -0.59544438123703,
+        0.37788620591163635,
+        -0.6611562967300415,
+        0.667895495891571,
+        0.31084170937538147,
+        -0.42228883504867554,
+        0.5119883418083191
+      ],
+      "std": [
+        0.043663933873176575,
+        0.0719655305147171,
+        0.06168781593441963,
+        0.05646670609712601,
+        0.24672336876392365,
+        0.0382271967828272,
+        0.4340234696865082,
+        0.08751874417066574,
+        0.28798049688339233,
+        0.49981093406677246
+      ],
+      "max": [
+        0.4219360053539276,
+        -0.17201299965381622,
+        0.33353498578071594,
+        -0.34213387966156006,
+        0.6040946245193481,
+        -0.5470927953720093,
+        0.9302739500999451,
+        0.70821213722229,
+        0.611838161945343,
+        1.0
+      ],
+      "min": [
+        0.12113799899816513,
+        -0.5658149719238281,
+        0.011877999641001225,
+        -0.745457112789154,
+        -0.5488103032112122,
+        -0.7786615490913391,
+        -0.8129586577415466,
+        -0.03098832629621029,
+        -0.6885839700698853,
+        0.0
+      ],
+      "q01": [
+        0.15309083521366118,
+        -0.5302998530864715,
+        0.01954999938607216,
+        -0.7032991063594818,
+        -0.47749382197856904,
+        -0.7447073876857757,
+        -0.7802488780021668,
+        0.07434731468558312,
+        -0.6129826271533966,
+        0.0
+      ],
+      "q99": [
+        0.38105886042118076,
+        -0.27944067239761344,
+        0.27689720153808595,
+        -0.4402932196855544,
+        0.5639739322662355,
+        -0.5716438913345336,
+        0.8863004243373871,
+        0.5411446142196659,
+        0.5678422379493714,
+        1.0
+      ]
+    },
+    "num_transitions": 65939,
+    "num_trajectories": 525
+  }
+}

fastumi_pickandplace_qwenPI_329v4/final_model/pytorch_model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9100766b6882b860c456af520cf4b9ec2281826a24b29db27f01b0e90275bb38
+size 12444501852

fastumi_pickandplace_qwenPI_329v4/summary.jsonl

@@ -0,0 +1,2 @@
+{"steps": 5000}
+{"steps": 10000}