Hugging Face's logo

kasohrab
/
theworld-spatial-rgpt

TensorBoard
Safetensors
the_world
Model card Files
xet
Metrics Community
kasohrab commited on
Commit
a191b83
·
verified ·
1 Parent(s): fef8717

Training in progress, step 1000

Browse files
Files changed (11) hide show
  1. config.json +103 -0
  2. logs/events.out.tfevents.1762186246.atl1-1-03-018-2-0.pace.gatech.edu.3161160.0 +3 -0
  3. logs/events.out.tfevents.1762203795.atl1-1-03-017-2-0.pace.gatech.edu.524460.0 +3 -0
  4. logs/events.out.tfevents.1762204453.atl1-1-03-017-2-0.pace.gatech.edu.536321.0 +3 -0
  5. logs/events.out.tfevents.1762205207.atl1-1-03-017-2-0.pace.gatech.edu.562857.0 +3 -0
  6. logs/events.out.tfevents.1762205788.atl1-1-03-017-2-0.pace.gatech.edu.576337.0 +3 -0
  7. logs/events.out.tfevents.1762206381.atl1-1-03-017-2-0.pace.gatech.edu.598273.0 +3 -0
  8. logs/events.out.tfevents.1762206877.atl1-1-03-017-2-0.pace.gatech.edu.606575.0 +3 -0
  9. logs/events.out.tfevents.1762208384.atl1-1-03-017-2-0.pace.gatech.edu.627879.0 +3 -0
  10. model.safetensors +3 -0
  11. training_args.bin +3 -0
config.json ADDED
@@ -0,0 +1,103 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Gemma3ForConditionalGeneration"
4
+ ],
5
+ "boi_token_index": 255999,
6
+ "cosmos_model_name": "nvidia/Cosmos-Predict2-2B-Video2World",
7
+ "dtype": "bfloat16",
8
+ "enable_world": true,
9
+ "eoi_token_index": 256000,
10
+ "eos_token_id": [
11
+ 1,
12
+ 106
13
+ ],
14
+ "freeze_cosmos_vae": true,
15
+ "freeze_gemma_language": true,
16
+ "freeze_gemma_vision": true,
17
+ "gemma_model_name": "google/gemma-3-4b-it",
18
+ "image_token_index": 262144,
19
+ "initializer_range": 0.02,
20
+ "mm_tokens_per_image": 256,
21
+ "model_type": "the_world",
22
+ "text_config": {
23
+ "_sliding_window_pattern": 6,
24
+ "attention_bias": false,
25
+ "attention_dropout": 0.0,
26
+ "attn_logit_softcapping": null,
27
+ "dtype": "bfloat16",
28
+ "final_logit_softcapping": null,
29
+ "head_dim": 256,
30
+ "hidden_activation": "gelu_pytorch_tanh",
31
+ "hidden_size": 2560,
32
+ "initializer_range": 0.02,
33
+ "intermediate_size": 10240,
34
+ "layer_types": [
35
+ "sliding_attention",
36
+ "sliding_attention",
37
+ "sliding_attention",
38
+ "sliding_attention",
39
+ "sliding_attention",
40
+ "full_attention",
41
+ "sliding_attention",
42
+ "sliding_attention",
43
+ "sliding_attention",
44
+ "sliding_attention",
45
+ "sliding_attention",
46
+ "full_attention",
47
+ "sliding_attention",
48
+ "sliding_attention",
49
+ "sliding_attention",
50
+ "sliding_attention",
51
+ "sliding_attention",
52
+ "full_attention",
53
+ "sliding_attention",
54
+ "sliding_attention",
55
+ "sliding_attention",
56
+ "sliding_attention",
57
+ "sliding_attention",
58
+ "full_attention",
59
+ "sliding_attention",
60
+ "sliding_attention",
61
+ "sliding_attention",
62
+ "sliding_attention",
63
+ "sliding_attention",
64
+ "full_attention",
65
+ "sliding_attention",
66
+ "sliding_attention",
67
+ "sliding_attention",
68
+ "sliding_attention"
69
+ ],
70
+ "max_position_embeddings": 131072,
71
+ "model_type": "gemma3_text",
72
+ "num_attention_heads": 8,
73
+ "num_hidden_layers": 34,
74
+ "num_key_value_heads": 4,
75
+ "query_pre_attn_scalar": 256,
76
+ "rms_norm_eps": 1e-06,
77
+ "rope_local_base_freq": 10000.0,
78
+ "rope_scaling": {
79
+ "factor": 8.0,
80
+ "rope_type": "linear"
81
+ },
82
+ "rope_theta": 1000000.0,
83
+ "sliding_window": 1024,
84
+ "use_cache": true,
85
+ "vocab_size": 262147
86
+ },
87
+ "transformers_version": "4.56.2",
88
+ "vision_config": {
89
+ "attention_dropout": 0.0,
90
+ "dtype": "bfloat16",
91
+ "hidden_act": "gelu_pytorch_tanh",
92
+ "hidden_size": 1152,
93
+ "image_size": 896,
94
+ "intermediate_size": 4304,
95
+ "layer_norm_eps": 1e-06,
96
+ "model_type": "siglip_vision_model",
97
+ "num_attention_heads": 16,
98
+ "num_channels": 3,
99
+ "num_hidden_layers": 27,
100
+ "patch_size": 14,
101
+ "vision_use_head": false
102
+ }
103
+ }
logs/events.out.tfevents.1762186246.atl1-1-03-018-2-0.pace.gatech.edu.3161160.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7471e4b68a0bd0f74d61d8894b3398e6b8e47a7304be6d790280910d1904b4a5
3
+ size 7092
logs/events.out.tfevents.1762203795.atl1-1-03-017-2-0.pace.gatech.edu.524460.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9148d9109e8ead4aee830f9e497126f1f1a9f7326f80e8c570aa355a41ed5e5e
3
+ size 7092
logs/events.out.tfevents.1762204453.atl1-1-03-017-2-0.pace.gatech.edu.536321.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dd99958c9f8a882d4e2747f76c245c3a40172e2874b189ad99dea9350618790c
3
+ size 7092
logs/events.out.tfevents.1762205207.atl1-1-03-017-2-0.pace.gatech.edu.562857.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:813fcd52434ff3f4638c4a479d99fd24b7e737fdb7ebec635758abc7886d94c3
3
+ size 7092
logs/events.out.tfevents.1762205788.atl1-1-03-017-2-0.pace.gatech.edu.576337.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:613ff517d03c889d8f8113b4bd77c391bb3dbf9f435b3ed2913828fdcd957e5c
3
+ size 7092
logs/events.out.tfevents.1762206381.atl1-1-03-017-2-0.pace.gatech.edu.598273.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:63d383e1d69d55b58bcd519b6a8ebef53ec51382c0a9f91fd04ed641a5f6209e
3
+ size 7092
logs/events.out.tfevents.1762206877.atl1-1-03-017-2-0.pace.gatech.edu.606575.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4aae5a9f65037f7915a8295412d59f9e9118a67fe5b7c777d4312cafb8d7b165
3
+ size 7092
logs/events.out.tfevents.1762208384.atl1-1-03-017-2-0.pace.gatech.edu.627879.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:60c14a8e604f828e4a731f370ca4ec3bf58b717d110d3984d425aed4b3d91867
3
+ size 11304
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:453664e11a600420e3e3d5eaaac60bf26998fd4d66a613d21302cd29f9730757
3
+ size 165707758
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:029703f4b98453fc2efb1da51c990842f74ddefbbe68be4b06472b09c23c4497
3
+ size 5432