jakegrigsby commited on
Commit
02e42b9
·
verified ·
1 Parent(s): 3ab95e5

Upload folder using huggingface_hub

Browse files
large-il/ckpts/policy_weights/policy_epoch_0.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5fc27c331463b84bfdb71bba2e9fcb46f374d411f55ab9b9b0cd0b494dade8a4
3
+ size 777764502
large-il/ckpts/policy_weights/policy_epoch_10.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:713e069f2bc610be8e1f1ceff07d3d5b26fc71c2df29e4be3acb0574921291ac
3
+ size 777764866
large-il/ckpts/policy_weights/policy_epoch_12.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b72c43c31e3507305eb3b54038c26c7b354ff6f1bec1f524558474cd0326a5a5
3
+ size 777764866
large-il/ckpts/policy_weights/policy_epoch_14.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:80a87cba4e420e014c06c80d3cd4ff7a0ab2efe7a698dd9b9c7547f7763a00cc
3
+ size 777764866
large-il/ckpts/policy_weights/policy_epoch_16.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:03cf87aef5c2774f65e9ab053af82d3a5f8b3e76e943e9b2de2763d105ad4f26
3
+ size 777764866
large-il/ckpts/policy_weights/policy_epoch_18.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ab3f2c1ac2a4595b3e722f03ccc965ed0c636b74ea3ce2e5b1ed2c2719e92c16
3
+ size 777764866
large-il/ckpts/policy_weights/policy_epoch_2.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f5c19122a074e35c5bc84c26b16ba9f62e6c262ad94b189a8fd61af53e02dcbc
3
+ size 777764502
large-il/ckpts/policy_weights/policy_epoch_20.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4e3f47f676e860177ae4c5e0e0c8f14a72a584dae5f29ee7c2405182692f9f9b
3
+ size 777764866
large-il/ckpts/policy_weights/policy_epoch_22.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e5e7feed04e865f26ac52c2bda6f0a4db3c7186a3f58ddd9ed3760154ff978d5
3
+ size 777764866
large-il/ckpts/policy_weights/policy_epoch_24.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1a2f1678ae5466a9251e8619ad200d6432b097b65acfca193690ab3b3c9733a1
3
+ size 777764866
large-il/ckpts/policy_weights/policy_epoch_26.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1c616dedfd0f195b00e33f186aab72d27e44a5e7a3f86b5b61a7f982b78cfbe2
3
+ size 777764866
large-il/ckpts/policy_weights/policy_epoch_28.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0fc6150dbd511a28456de74de146b452f106c4989d410d148f4c9cc4730377c6
3
+ size 777764866
large-il/ckpts/policy_weights/policy_epoch_30.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3e7706a3edd73b700fad43a384ba92e0e15237147d06fb2d8258adc71761ae3c
3
+ size 777764866
large-il/ckpts/policy_weights/policy_epoch_32.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:612da55f0e7cae1250a80658665a5c7fc2a0d15827b298f359f85317c42b902f
3
+ size 777764866
large-il/ckpts/policy_weights/policy_epoch_34.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6702e24f8a7faaf4bbb198f4ffddf5466315d94667c932ec29a69edb8c5e64d2
3
+ size 777764866
large-il/ckpts/policy_weights/policy_epoch_36.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e110f1e6552b6428175accfd5fe853e72f9b691aeb30548423b79e047e737464
3
+ size 777764866
large-il/ckpts/policy_weights/policy_epoch_38.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0e7037555b8b2ad3c449d87760ab784575b4976c728b7f45201ecdaea07f0d62
3
+ size 777764866
large-il/ckpts/policy_weights/policy_epoch_4.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b58bb4e9a51588f1a24939e888853f69875f52f46be4795d0fee71aac6a85632
3
+ size 777764502
large-il/ckpts/policy_weights/policy_epoch_40.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8a43fcc7798c41df404a72486002a2d6a7c3ef9f988b948ae062af6a2bb1e89b
3
+ size 777764866
large-il/ckpts/policy_weights/policy_epoch_42.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cadbfcb724f123ba1b0c4c151e1155e9b36d5940d55a986e2dd62f555e6e4316
3
+ size 777764866
large-il/ckpts/policy_weights/policy_epoch_44.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b82aeeb234d9f31c2bfee05477383191c8215c7b8f40bd97b7cb795fa11d737b
3
+ size 777764866
large-il/ckpts/policy_weights/policy_epoch_46.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9572f312324554ada12abe27e0480a3a125018297a87aa0f51ef6ab7b3dafc91
3
+ size 777764866
large-il/ckpts/policy_weights/policy_epoch_48.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6e3f25753755e0d9eeec9870231b6885139d3e93a6514cc5976684807e964ef6
3
+ size 777764866
large-il/ckpts/policy_weights/policy_epoch_50.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c1c31a454061c71e1e44365d28f0ef29b1fc30198d9e2b4c38b6299df97d4459
3
+ size 777764866
large-il/ckpts/policy_weights/policy_epoch_6.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5365e038780b74ee6997b2f3d3aeed08af84258013bf78b8948eb1a950084d8f
3
+ size 777764502
large-il/ckpts/policy_weights/policy_epoch_8.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dd12e66bc6474c8f36a37369a009035b19a9a09d44334d8ab5b5bbc358fdef9d
3
+ size 777764502
large-il/config.txt ADDED
@@ -0,0 +1,106 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Parameters for Actor:
2
+ # ==============================================================================
3
+ Actor.activation = 'leaky_relu'
4
+ Actor.cont_dist_kind = 'normal'
5
+ Actor.d_hidden = 512
6
+ Actor.dropout_p = 0.0
7
+ Actor.gmm_modes = 5
8
+ Actor.log_std_high = 2.0
9
+ Actor.log_std_low = -5.0
10
+ Actor.n_layers = 2
11
+
12
+ # Parameters for Agent:
13
+ # ==============================================================================
14
+ Agent.fake_filter = True
15
+ Agent.gamma = 0.999
16
+ Agent.num_critics = 4
17
+ Agent.num_critics_td = 2
18
+ Agent.offline_coeff = 1.0
19
+ Agent.online_coeff = 0.0
20
+ Agent.popart = True
21
+ Agent.reward_multiplier = 10.0
22
+ Agent.tau = 0.003
23
+ Agent.use_multigamma = False
24
+ Agent.use_target_actor = True
25
+
26
+ # Parameters for Experiment:
27
+ # ==============================================================================
28
+ Experiment.batches_per_update = 1
29
+ Experiment.critic_loss_weight = 10.0
30
+ Experiment.env_mode = 'async'
31
+ Experiment.force_reset_train_envs_every = None
32
+ Experiment.grad_clip = 1.0
33
+ Experiment.has_replay_buffer_rights = True
34
+ Experiment.l2_coeff = 0.001
35
+ Experiment.learning_rate = 0.0001
36
+ Experiment.local_time_optimizer = False
37
+ Experiment.lr_warmup_steps = 500
38
+ Experiment.mixed_precision = 'no'
39
+ Experiment.padded_sampling = 'none'
40
+ Experiment.save_trajs_as = 'npz'
41
+ Experiment.stagger_traj_file_lengths = True
42
+ Experiment.wandb_group_name = None
43
+
44
+ # Parameters for FlashAttention:
45
+ # ==============================================================================
46
+ FlashAttention.window_size = (-1, -1)
47
+
48
+ # Parameters for MetamonTstepEncoder:
49
+ # ==============================================================================
50
+ MetamonTstepEncoder.d_model = 160
51
+ MetamonTstepEncoder.extra_emb_dim = 18
52
+ MetamonTstepEncoder.n_heads = 8
53
+ MetamonTstepEncoder.n_layers = 5
54
+ MetamonTstepEncoder.scratch_tokens = 11
55
+ MetamonTstepEncoder.token_mask_aug = False
56
+
57
+ # Parameters for Multigammas:
58
+ # ==============================================================================
59
+ Multigammas.continuous = [0.1, 0.9, 0.95, 0.97, 0.99, 0.995]
60
+ Multigammas.discrete = [0.1, 0.9, 0.95, 0.97, 0.99, 0.995]
61
+
62
+ # Parameters for MultiModalEmbedding:
63
+ # ==============================================================================
64
+ MultiModalEmbedding.dropout = 0.05
65
+ MultiModalEmbedding.numerical_tokens = 6
66
+
67
+ # Parameters for NCritics:
68
+ # ==============================================================================
69
+ NCritics.activation = 'leaky_relu'
70
+ NCritics.d_hidden = 512
71
+ NCritics.dropout_p = 0.0
72
+ NCritics.n_layers = 2
73
+
74
+ # Parameters for PopArtLayer:
75
+ # ==============================================================================
76
+ PopArtLayer.beta = 0.0005
77
+ PopArtLayer.init_nu = 100.0
78
+
79
+ # Parameters for TformerTrajEncoder:
80
+ # ==============================================================================
81
+ TformerTrajEncoder.activation = 'leaky_relu'
82
+ TformerTrajEncoder.causal = True
83
+ TformerTrajEncoder.d_ff = 5120
84
+ TformerTrajEncoder.d_model = 1280
85
+ TformerTrajEncoder.dropout_attn = 0.0
86
+ TformerTrajEncoder.dropout_emb = 0.05
87
+ TformerTrajEncoder.dropout_ff = 0.05
88
+ TformerTrajEncoder.dropout_qkv = 0.0
89
+ TformerTrajEncoder.head_scaling = True
90
+ TformerTrajEncoder.n_heads = 20
91
+ TformerTrajEncoder.n_layers = 9
92
+ TformerTrajEncoder.norm = 'layer'
93
+ TformerTrajEncoder.normformer_norms = True
94
+ TformerTrajEncoder.sigma_reparam = True
95
+
96
+ # Parameters for TimestepTransformer:
97
+ # ==============================================================================
98
+ # None.
99
+
100
+ # Parameters for TokenEmbedding:
101
+ # ==============================================================================
102
+ # None.
103
+
104
+ # Parameters for TransformerTurnEmbedding:
105
+ # ==============================================================================
106
+ TransformerTurnEmbedding.dropout = 0.05