Hugging Face's logo

zeeshaan-ai
/
GetSoloTech

Text Generation
Transformers
Safetensors
qwen3
solo
fine-tuned
lora
unsloth
conversational
text-generation-inference
Model card Files
xet
 Community
zeeshaan-ai commited on
Commit
cf2a589
·
verified ·
1 Parent(s): 1b32a66

Upload model

Browse files
Files changed (5) hide show
  1. config.json +68 -59
  2. model.safetensors +2 -2
  3. policy_postprocessor.json +2 -2
  4. policy_preprocessor.json +10 -25
  5. train_config.json +34 -53
config.json CHANGED
@@ -1,61 +1,70 @@
1
  {
2
- "architectures": [
3
- "Qwen3ForCausalLM"
4
- ],
5
- "attention_bias": false,
6
- "attention_dropout": 0.0,
7
- "torch_dtype": "bfloat16",
8
- "eos_token_id": 151645,
9
- "head_dim": 128,
10
- "hidden_act": "silu",
11
- "hidden_size": 1024,
12
- "initializer_range": 0.02,
13
- "intermediate_size": 3072,
14
- "layer_types": [
15
- "full_attention",
16
- "full_attention",
17
- "full_attention",
18
- "full_attention",
19
- "full_attention",
20
- "full_attention",
21
- "full_attention",
22
- "full_attention",
23
- "full_attention",
24
- "full_attention",
25
- "full_attention",
26
- "full_attention",
27
- "full_attention",
28
- "full_attention",
29
- "full_attention",
30
- "full_attention",
31
- "full_attention",
32
- "full_attention",
33
- "full_attention",
34
- "full_attention",
35
- "full_attention",
36
- "full_attention",
37
- "full_attention",
38
- "full_attention",
39
- "full_attention",
40
- "full_attention",
41
- "full_attention",
42
- "full_attention"
43
- ],
44
- "max_position_embeddings": 40960,
45
- "max_window_layers": 28,
46
- "model_type": "qwen3",
47
- "num_attention_heads": 16,
48
- "num_hidden_layers": 28,
49
- "num_key_value_heads": 8,
50
- "pad_token_id": 151654,
51
- "rms_norm_eps": 1e-06,
52
- "rope_scaling": null,
53
- "rope_theta": 1000000,
54
- "sliding_window": null,
55
- "tie_word_embeddings": true,
56
- "unsloth_fixed": true,
57
- "unsloth_version": "2026.2.1",
58
- "use_cache": true,
59
- "use_sliding_window": false,
60
- "vocab_size": 151936
 
 
 
 
 
 
 
 
 
61
  }
 
1
  {
2
+ "type": "act",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.state": {
6
+ "type": "STATE",
7
+ "shape": [
8
+ 6
9
+ ]
10
+ },
11
+ "observation.images.side": {
12
+ "type": "VISUAL",
13
+ "shape": [
14
+ 3,
15
+ 1080,
16
+ 1920
17
+ ]
18
+ },
19
+ "observation.images.front": {
20
+ "type": "VISUAL",
21
+ "shape": [
22
+ 3,
23
+ 480,
24
+ 640
25
+ ]
26
+ }
27
+ },
28
+ "output_features": {
29
+ "action": {
30
+ "type": "ACTION",
31
+ "shape": [
32
+ 6
33
+ ]
34
+ }
35
+ },
36
+ "device": "cuda",
37
+ "use_amp": false,
38
+ "push_to_hub": false,
39
+ "repo_id": null,
40
+ "private": null,
41
+ "tags": null,
42
+ "license": null,
43
+ "pretrained_path": null,
44
+ "chunk_size": 100,
45
+ "n_action_steps": 100,
46
+ "normalization_mapping": {
47
+ "VISUAL": "MEAN_STD",
48
+ "STATE": "MEAN_STD",
49
+ "ACTION": "MEAN_STD"
50
+ },
51
+ "vision_backbone": "resnet18",
52
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
53
+ "replace_final_stride_with_dilation": false,
54
+ "pre_norm": false,
55
+ "dim_model": 512,
56
+ "n_heads": 8,
57
+ "dim_feedforward": 3200,
58
+ "feedforward_activation": "relu",
59
+ "n_encoder_layers": 4,
60
+ "n_decoder_layers": 1,
61
+ "use_vae": true,
62
+ "latent_dim": 32,
63
+ "n_vae_encoder_layers": 4,
64
+ "temporal_ensemble_coeff": null,
65
+ "dropout": 0.1,
66
+ "kl_weight": 10.0,
67
+ "optimizer_lr": 1e-05,
68
+ "optimizer_weight_decay": 0.0001,
69
+ "optimizer_lr_backbone": 1e-05
70
  }
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3139913f522617464c0ccb05c83af97cfc2b9517dbfb2e58f361f5de3a4b7d85
3
- size 1192135096
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b0df965282f8816da215012b5b1efbfbfe86427c06e548b9ff4755ae76d99a84
3
+ size 206699736
policy_postprocessor.json CHANGED
@@ -14,9 +14,9 @@
14
  }
15
  },
16
  "norm_map": {
17
- "ACTION": "MEAN_STD",
18
  "STATE": "MEAN_STD",
19
- "VISUAL": "IDENTITY"
20
  }
21
  },
22
  "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors"
 
14
  }
15
  },
16
  "norm_map": {
17
+ "VISUAL": "MEAN_STD",
18
  "STATE": "MEAN_STD",
19
+ "ACTION": "MEAN_STD"
20
  }
21
  },
22
  "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors"
policy_preprocessor.json CHANGED
@@ -11,6 +11,13 @@
11
  "registry_name": "to_batch_processor",
12
  "config": {}
13
  },
 
 
 
 
 
 
 
14
  {
15
  "registry_name": "normalizer_processor",
16
  "config": {
@@ -46,34 +53,12 @@
46
  }
47
  },
48
  "norm_map": {
49
- "ACTION": "MEAN_STD",
50
  "STATE": "MEAN_STD",
51
- "VISUAL": "IDENTITY"
52
  }
53
  },
54
- "state_file": "policy_preprocessor_step_2_normalizer_processor.safetensors"
55
- },
56
- {
57
- "registry_name": "pi05_prepare_state_tokenizer_processor_step",
58
- "config": {}
59
- },
60
- {
61
- "registry_name": "tokenizer_processor",
62
- "config": {
63
- "max_length": 200,
64
- "task_key": "task",
65
- "padding_side": "right",
66
- "padding": "max_length",
67
- "truncation": true,
68
- "tokenizer_name": "google/paligemma-3b-pt-224"
69
- }
70
- },
71
- {
72
- "registry_name": "device_processor",
73
- "config": {
74
- "device": "cuda",
75
- "float_dtype": null
76
- }
77
  }
78
  ]
79
  }
 
11
  "registry_name": "to_batch_processor",
12
  "config": {}
13
  },
14
+ {
15
+ "registry_name": "device_processor",
16
+ "config": {
17
+ "device": "cuda",
18
+ "float_dtype": null
19
+ }
20
+ },
21
  {
22
  "registry_name": "normalizer_processor",
23
  "config": {
 
53
  }
54
  },
55
  "norm_map": {
56
+ "VISUAL": "MEAN_STD",
57
  "STATE": "MEAN_STD",
58
+ "ACTION": "MEAN_STD"
59
  }
60
  },
61
+ "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
62
  }
63
  ]
64
  }
train_config.json CHANGED
@@ -81,7 +81,7 @@
81
  },
82
  "env": null,
83
  "policy": {
84
- "type": "pi05",
85
  "n_obs_steps": 1,
86
  "input_features": {
87
  "observation.state": {
@@ -117,54 +117,41 @@
117
  },
118
  "device": "cuda",
119
  "use_amp": false,
120
- "push_to_hub": true,
121
- "repo_id": "zeeshaan-ai/GetSoloTech",
122
  "private": null,
123
  "tags": null,
124
  "license": null,
125
- "pretrained_path": "lerobot/pi05_base",
126
- "paligemma_variant": "gemma_2b",
127
- "action_expert_variant": "gemma_300m",
128
- "dtype": "bfloat16",
129
- "chunk_size": 50,
130
- "n_action_steps": 50,
131
- "max_state_dim": 32,
132
- "max_action_dim": 32,
133
- "num_inference_steps": 10,
134
- "time_sampling_beta_alpha": 1.5,
135
- "time_sampling_beta_beta": 1.0,
136
- "time_sampling_scale": 0.999,
137
- "time_sampling_offset": 0.001,
138
- "min_period": 0.004,
139
- "max_period": 4.0,
140
- "image_resolution": [
141
- 224,
142
- 224
143
- ],
144
- "empty_cameras": 0,
145
- "tokenizer_max_length": 200,
146
  "normalization_mapping": {
147
- "ACTION": "MEAN_STD",
148
  "STATE": "MEAN_STD",
149
- "VISUAL": "IDENTITY"
150
  },
151
- "gradient_checkpointing": true,
152
- "compile_model": true,
153
- "compile_mode": "max-autotune",
154
- "optimizer_lr": 2.5e-05,
155
- "optimizer_betas": [
156
- 0.9,
157
- 0.95
158
- ],
159
- "optimizer_eps": 1e-08,
160
- "optimizer_weight_decay": 0.01,
161
- "optimizer_grad_clip_norm": 1.0,
162
- "scheduler_warmup_steps": 1000,
163
- "scheduler_decay_steps": 30000,
164
- "scheduler_decay_lr": 2.5e-06
 
 
 
 
 
165
  },
166
  "output_dir": "/workspace/outputs",
167
- "job_name": "qqqqqqqqqqqqqqqqqvlaaa",
168
  "resume": false,
169
  "seed": 1000,
170
  "num_workers": 4,
@@ -173,26 +160,20 @@
173
  "eval_freq": 20000,
174
  "log_freq": 200,
175
  "save_checkpoint": true,
176
- "save_freq": 1000,
177
  "use_policy_training_preset": true,
178
  "optimizer": {
179
  "type": "adamw",
180
- "lr": 2.5e-05,
181
- "weight_decay": 0.01,
182
- "grad_clip_norm": 1.0,
183
  "betas": [
184
  0.9,
185
- 0.95
186
  ],
187
  "eps": 1e-08
188
  },
189
- "scheduler": {
190
- "type": "cosine_decay_with_warmup",
191
- "num_warmup_steps": 1000,
192
- "num_decay_steps": 30000,
193
- "peak_lr": 2.5e-05,
194
- "decay_lr": 2.5e-06
195
- },
196
  "eval": {
197
  "n_episodes": 50,
198
  "batch_size": 50,
 
81
  },
82
  "env": null,
83
  "policy": {
84
+ "type": "act",
85
  "n_obs_steps": 1,
86
  "input_features": {
87
  "observation.state": {
 
117
  },
118
  "device": "cuda",
119
  "use_amp": false,
120
+ "push_to_hub": false,
121
+ "repo_id": null,
122
  "private": null,
123
  "tags": null,
124
  "license": null,
125
+ "pretrained_path": null,
126
+ "chunk_size": 100,
127
+ "n_action_steps": 100,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
128
  "normalization_mapping": {
129
+ "VISUAL": "MEAN_STD",
130
  "STATE": "MEAN_STD",
131
+ "ACTION": "MEAN_STD"
132
  },
133
+ "vision_backbone": "resnet18",
134
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
135
+ "replace_final_stride_with_dilation": false,
136
+ "pre_norm": false,
137
+ "dim_model": 512,
138
+ "n_heads": 8,
139
+ "dim_feedforward": 3200,
140
+ "feedforward_activation": "relu",
141
+ "n_encoder_layers": 4,
142
+ "n_decoder_layers": 1,
143
+ "use_vae": true,
144
+ "latent_dim": 32,
145
+ "n_vae_encoder_layers": 4,
146
+ "temporal_ensemble_coeff": null,
147
+ "dropout": 0.1,
148
+ "kl_weight": 10.0,
149
+ "optimizer_lr": 1e-05,
150
+ "optimizer_weight_decay": 0.0001,
151
+ "optimizer_lr_backbone": 1e-05
152
  },
153
  "output_dir": "/workspace/outputs",
154
+ "job_name": "vlaaaamodeltest",
155
  "resume": false,
156
  "seed": 1000,
157
  "num_workers": 4,
 
160
  "eval_freq": 20000,
161
  "log_freq": 200,
162
  "save_checkpoint": true,
163
+ "save_freq": 10000,
164
  "use_policy_training_preset": true,
165
  "optimizer": {
166
  "type": "adamw",
167
+ "lr": 1e-05,
168
+ "weight_decay": 0.0001,
169
+ "grad_clip_norm": 10.0,
170
  "betas": [
171
  0.9,
172
+ 0.999
173
  ],
174
  "eps": 1e-08
175
  },
176
+ "scheduler": null,
 
 
 
 
 
 
177
  "eval": {
178
  "n_episodes": 50,
179
  "batch_size": 50,