Hugging Face's logo

xixibuxixi
/
plot-checkpoints

Safetensors
Model card Files
xet
 Community
plot-checkpoints
File size: 4,648 Bytes
45c6d49
 
f8e1b92
45c6d49
 
 
 
 
 
 
 
 
 
 
 
 
f8e1b92
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2b2d8ae
 
 
 
 
 
 
 
 
 
45c6d49
 
f8e1b92
 
 
 
 
 
 
 
45c6d49
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f8e1b92
45c6d49
 
 
 
 
 
 
f8e1b92
 
 
 
 
 
 
2b2d8ae
 
45c6d49
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
{
  "schema_version": "plot-selected-checkpoints-v1",
  "updated_utc": "2026-09-20",
  "checkpoints": [
    {
      "path": "m1/m1_flow_2nodes_b32_eval10k_v2/step_001476000.pt",
      "role": "selected M1 operational resume checkpoint",
      "bytes": 5491811243,
      "sha256": "8f75b7a43f049a5b81b8133b8374749b5915cc1494d4c3b9d14574f9b2073048",
      "source": "legacy Polis run m1_flow_2nodes_b32_eval10k_v2",
      "step": 1476000,
      "code": "legacy source hashes; compatible PLOT migration at 4b6ce14",
      "target_worktree_commit": "639ee45869e45cc1b943f1e206ee5f74fc261eb6",
      "validation": "torch.load with mmap succeeded; current MultiViewVoxelDiT strict-load on zhizhou-avgen-2 matched all 288 tensors; AdamW restored all 288 optimizer states at lr 1e-4"
    },
    {
      "path": "m3/m3_simple_fulltrain_player_flow_step7365/step_0007365.pt",
      "role": "selected current M3 exploration/resume checkpoint",
      "bytes": 5533183645,
      "sha256": "b598bfc041177d152841484df5e9e0ef5cf17594c70031d24316bf01b1aef80c",
      "step": 7365,
      "source": "m3_fulltrain_8gpu_b4_player_flow_20260919",
      "architecture": "M3 Simple: ROI appearance attention then channel concat with video/voxel/actor maps into joint patch embedder",
      "compatible_code_commit": "efbe000c1c5f8d93a103780b32707f6f90f6b7bb",
      "code_note": "This is the subsequent committed implementation, not a claim that the training launched at this commit.",
      "dataset_release": "polis_two_player_fixed_skins_complete_20260917_360p (repaired references)",
      "contains": [
        "model",
        "optimizer"
      ],
      "wandb": "https://wandb.ai/ckx23-tsinghua-university/plot-m3/runs/fhh63zsg",
      "loss": "full_flow + player_flow + 0.1 * player_RGB + 0.025 * player_edge",
      "batch": {
        "per_gpu": 4,
        "gpus": 8,
        "gradient_accumulation": 1,
        "effective": 32
      },
      "context_frames": 9,
      "condition_frames": 1,
      "rollout_block_frames": 8,
      "dependencies": {
        "pixel_vae_sha256": "eb634803c94aeea980046961382f2ab67157aa71e92c5a183a35e4f61f8cbc36"
      },
      "limitations": "Exploration checkpoint; not claimed to solve player appearance. Validation logs use a small fixed probe set."
    },
    {
      "path": "m2/m2_s02_s10_attack_v7/deployed_checkpoint.pt",
      "role": "complete M2 deployed model (edit + player + attack)",
      "step": 8000,
      "bytes": 86807075,
      "sha256": "8f7e85a948b380dd7c57446e73abb12d2768a2a828130fbfdf8589bd83dc1cc3",
      "validation": "Current Plot strict-load: 375 tensors; all source player tensors unchanged.",
      "compatibility_note": "M2 and M3 item ID 12 differ; see TRAINING_ASSETS.md.",
      "compatible_code_commit": "efbe000c1c5f8d93a103780b32707f6f90f6b7bb"
    }
  ],
  "publishing": {
    "hugging_face_repo": "xixibuxixi/plot-checkpoints",
    "private": true,
    "status": "uploaded and verified",
    "artifacts_revision": "d8691047695ab0eebdb7317962f88e7cd392df51",
    "verification": "All five new artifact sizes and SHA-256 values matched local files."
  },
  "historical_dependencies": {
    "m3_validated_train_index": {
      "path": "derived/m3/validated/train_c65.pt",
      "bytes": 344160441,
      "sha256": "d0e483e85949bccdca3899c0c13c7d95e35f8a6b8dc3662aa8aa20d504521b7a"
    },
    "m3_validated_val_index": {
      "path": "derived/m3/validated/val_id_c65.pt",
      "bytes": 21900351,
      "sha256": "6fdad62f55921ca5b53bc45c35d564753abfb28d646f4062b43c5dee43133399"
    },
    "block_vocabulary": {
      "path": "derived/common/block_vocabulary.json",
      "sha256": "ee8e3c89f810d2f6b49a1e07ba168ba187b85623eea168aaa324e8196060c6de",
      "validation": "shared-cluster and historical-H100 copies match"
    }
  },
  "historical_dataset": {
    "repository": "xixibuxixi/polis-v1",
    "local_release": "polis_v1_20260909_360p",
    "release_json_sha256": "d3a69316575d4e87d91a991a6fe1fadaec0baf7ab31d26f67b782b3f077ad8ba",
    "train_jsonl_sha256": "47190499ec76ff24bbca5600ca4f1bd39fafec9b87618ac65039bf0109cb95f6",
    "val_id_jsonl_sha256": "c2e2646dfa52a93cc8d15796e2d0019ffbf14901ead064520dc631c3e4c74b6b",
    "validation": "shared-cluster and historical-H100 copies match for all three files"
  },
  "retired_m3": {
    "paths": [
      "m3/m3_entity_reference_stage1/",
      "m3/player_identity_real_finetune/"
    ],
    "recoverable_revision": "9d51ae5bc6bf9ece0392f7b92b1ef29a33cb5ce9",
    "reason": "User selected current M3 Simple architecture and requested removal of these two obsolete assets."
  },
  "training_assets_manifest": "TRAINING_ASSETS.json"
}