kissin42's picture
Add PushBlock 28k policy bundle
22546a1 verified
Raw History Blame Contribute Delete
16 kB
{
"bundle_format_version": 2,
"package": "causal-gpt-rl",
"package_version": "0.16.0",
"model_config": {
"network_name": "Llama",
"num_layers": 4,
"d_model": 256,
"num_heads": 8,
"dropout": 0.05,
"use_eos": true,
"use_bos_action_gate": true,
"rope_theta": 1000.0,
"intermediate_size": 1024,
"max_position_embeddings": 256,
"context_length": 32
},
"state_specs": [
{
"type": "continuous",
"size": 105,
"dtype": "float32",
"low": [
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null
],
"high": [
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null
],
"squash": null
},
{
"type": "continuous",
"size": 105,
"dtype": "float32",
"low": [
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null
],
"high": [
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null,
null
],
"squash": null
}
],
"action_specs": [
{
"type": "discrete",
"size": 7,
"dtype": "int64",
"low": [
null,
null,
null,
null,
null,
null,
null
],
"high": [
null,
null,
null,
null,
null,
null,
null
],
"squash": null
}
],
"context_length": 32,
"requires_capabilities": [
"hybrid_state"
],
"state_container": {
"type": "Tuple",
"spaces": [
{
"type": "Box",
"shape": [
105
],
"dtype": "float32",
"low": [
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf"
],
"high": [
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf"
]
},
{
"type": "Box",
"shape": [
105
],
"dtype": "float32",
"low": [
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf",
"-inf"
],
"high": [
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf",
"inf"
]
}
]
},
"action_container": {
"type": "Discrete",
"n": 7,
"start": 0
},
"state_normalization": {
"embedded": true,
"legacy_sidecar": false
},
"env_id": "PushBlock",
"serving": {
"bos_cache_mode": "discard"
}
}