Training in progress, step 70000, checkpoint
Browse files
last-checkpoint/generation_config.json
CHANGED
|
@@ -1,5 +1,4 @@
|
|
| 1 |
{
|
| 2 |
-
"_from_model_config": true,
|
| 3 |
"bos_token_id": 0,
|
| 4 |
"do_sample": true,
|
| 5 |
"eos_token_id": 2,
|
|
|
|
| 1 |
{
|
|
|
|
| 2 |
"bos_token_id": 0,
|
| 3 |
"do_sample": true,
|
| 4 |
"eos_token_id": 2,
|
last-checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 409608164
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3ecb5e044c918d954c91a467f251821ada8da874b0d0ac17a9576f80c2c3e478
|
| 3 |
size 409608164
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 814647162
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:505e9b671aa22818b29b43d0cbe34b09795f1ea393cc430ee3f77aa6a1eca9fa
|
| 3 |
size 814647162
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:77dc4e293062bfce5492775624cabb52f7debf0552a3b6d02eb1e46a77a45612
|
| 3 |
size 14244
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d7b405293c7c61627577f0ffca1ebd7a6cc00745b4ef6a4df74ea0a9189fd0ef
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "/mimer/NOBACKUP/groups/naiss2023-6-290/stefano/models//PROTAC-Splitter-EncoderDecoder-lr_cosine
|
| 4 |
-
"epoch": 6.
|
| 5 |
"eval_steps": 2500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -1043,6 +1043,118 @@
|
|
| 1043 |
"learning_rate": 1.3684575241351546e-05,
|
| 1044 |
"loss": 0.0003,
|
| 1045 |
"step": 65000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1046 |
}
|
| 1047 |
],
|
| 1048 |
"logging_steps": 500,
|
|
@@ -1062,7 +1174,7 @@
|
|
| 1062 |
"attributes": {}
|
| 1063 |
}
|
| 1064 |
},
|
| 1065 |
-
"total_flos":
|
| 1066 |
"train_batch_size": 128,
|
| 1067 |
"trial_name": null,
|
| 1068 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.5963172804532578,
|
| 3 |
+
"best_model_checkpoint": "/mimer/NOBACKUP/groups/naiss2023-6-290/stefano/models//PROTAC-Splitter-EncoderDecoder-lr_cosine/checkpoint-70000",
|
| 4 |
+
"epoch": 6.904714933912015,
|
| 5 |
"eval_steps": 2500,
|
| 6 |
+
"global_step": 70000,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1043 |
"learning_rate": 1.3684575241351546e-05,
|
| 1044 |
"loss": 0.0003,
|
| 1045 |
"step": 65000
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 6.460840402446242,
|
| 1049 |
+
"grad_norm": 0.01365108322352171,
|
| 1050 |
+
"learning_rate": 1.3335466168024479e-05,
|
| 1051 |
+
"loss": 0.0003,
|
| 1052 |
+
"step": 65500
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 6.510159794831328,
|
| 1056 |
+
"grad_norm": 0.01254518423229456,
|
| 1057 |
+
"learning_rate": 1.2989240908505564e-05,
|
| 1058 |
+
"loss": 0.0003,
|
| 1059 |
+
"step": 66000
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 6.5594791872164135,
|
| 1063 |
+
"grad_norm": 0.0054794480092823505,
|
| 1064 |
+
"learning_rate": 1.2645985059801791e-05,
|
| 1065 |
+
"loss": 0.0003,
|
| 1066 |
+
"step": 66500
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 6.608798579601499,
|
| 1070 |
+
"grad_norm": 0.007447356358170509,
|
| 1071 |
+
"learning_rate": 1.2306460783748302e-05,
|
| 1072 |
+
"loss": 0.0003,
|
| 1073 |
+
"step": 67000
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 6.658117971986585,
|
| 1077 |
+
"grad_norm": 0.0031654227059334517,
|
| 1078 |
+
"learning_rate": 1.196939123002433e-05,
|
| 1079 |
+
"loss": 0.0003,
|
| 1080 |
+
"step": 67500
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 6.707437364371671,
|
| 1084 |
+
"grad_norm": 0.009027605876326561,
|
| 1085 |
+
"learning_rate": 1.1635543223774156e-05,
|
| 1086 |
+
"loss": 0.0003,
|
| 1087 |
+
"step": 68000
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 6.756756756756757,
|
| 1091 |
+
"grad_norm": 0.008913296274840832,
|
| 1092 |
+
"learning_rate": 1.130499930198588e-05,
|
| 1093 |
+
"loss": 0.0003,
|
| 1094 |
+
"step": 68500
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 6.806076149141843,
|
| 1098 |
+
"grad_norm": 0.005660299211740494,
|
| 1099 |
+
"learning_rate": 1.0977841184781343e-05,
|
| 1100 |
+
"loss": 0.0003,
|
| 1101 |
+
"step": 69000
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 6.855395541526929,
|
| 1105 |
+
"grad_norm": 0.006413063500076532,
|
| 1106 |
+
"learning_rate": 1.0654793624981977e-05,
|
| 1107 |
+
"loss": 0.0003,
|
| 1108 |
+
"step": 69500
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 6.904714933912015,
|
| 1112 |
+
"grad_norm": 0.004667436238378286,
|
| 1113 |
+
"learning_rate": 1.0334641736320572e-05,
|
| 1114 |
+
"loss": 0.0003,
|
| 1115 |
+
"step": 70000
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 6.904714933912015,
|
| 1119 |
+
"eval_all_ligands_equal": 0.5963172804532578,
|
| 1120 |
+
"eval_e3_equal": 0.8271954674220963,
|
| 1121 |
+
"eval_e3_graph_edit_distance": Infinity,
|
| 1122 |
+
"eval_e3_graph_edit_distance_norm": Infinity,
|
| 1123 |
+
"eval_e3_has_attachment_point(s)": 0.9876062322946175,
|
| 1124 |
+
"eval_e3_heavy_atoms_difference": 0.32790368271954673,
|
| 1125 |
+
"eval_e3_heavy_atoms_difference_norm": 0.004086581188234727,
|
| 1126 |
+
"eval_e3_tanimoto_similarity": 0.0,
|
| 1127 |
+
"eval_e3_valid": 0.9876062322946175,
|
| 1128 |
+
"eval_has_all_attachment_points": 0.9836225212464589,
|
| 1129 |
+
"eval_has_three_substructures": 0.9996458923512748,
|
| 1130 |
+
"eval_heavy_atoms_difference": 6.53257790368272,
|
| 1131 |
+
"eval_heavy_atoms_difference_norm": 0.08691394560206064,
|
| 1132 |
+
"eval_linker_equal": 0.8436614730878187,
|
| 1133 |
+
"eval_linker_graph_edit_distance": Infinity,
|
| 1134 |
+
"eval_linker_graph_edit_distance_norm": Infinity,
|
| 1135 |
+
"eval_linker_has_attachment_point(s)": 0.9970786118980169,
|
| 1136 |
+
"eval_linker_heavy_atoms_difference": 0.2073300283286119,
|
| 1137 |
+
"eval_linker_heavy_atoms_difference_norm": -0.00028415548137912695,
|
| 1138 |
+
"eval_linker_tanimoto_similarity": 0.0,
|
| 1139 |
+
"eval_linker_valid": 0.9970786118980169,
|
| 1140 |
+
"eval_loss": 0.3754393458366394,
|
| 1141 |
+
"eval_num_fragments": 3.0,
|
| 1142 |
+
"eval_poi_equal": 0.7956798866855525,
|
| 1143 |
+
"eval_poi_graph_edit_distance": Infinity,
|
| 1144 |
+
"eval_poi_graph_edit_distance_norm": Infinity,
|
| 1145 |
+
"eval_poi_has_attachment_point(s)": 0.93546388101983,
|
| 1146 |
+
"eval_poi_heavy_atoms_difference": 1.9802584985835694,
|
| 1147 |
+
"eval_poi_heavy_atoms_difference_norm": 0.06374572149016507,
|
| 1148 |
+
"eval_poi_tanimoto_similarity": 0.0,
|
| 1149 |
+
"eval_poi_valid": 0.93546388101983,
|
| 1150 |
+
"eval_reassembly": 0.6042847025495751,
|
| 1151 |
+
"eval_reassembly_nostereo": 0.6316395184135978,
|
| 1152 |
+
"eval_runtime": 2293.8906,
|
| 1153 |
+
"eval_samples_per_second": 4.924,
|
| 1154 |
+
"eval_steps_per_second": 0.077,
|
| 1155 |
+
"eval_tanimoto_similarity": 0.0,
|
| 1156 |
+
"eval_valid": 0.921742209631728,
|
| 1157 |
+
"step": 70000
|
| 1158 |
}
|
| 1159 |
],
|
| 1160 |
"logging_steps": 500,
|
|
|
|
| 1174 |
"attributes": {}
|
| 1175 |
}
|
| 1176 |
},
|
| 1177 |
+
"total_flos": 6.06433214147116e+17,
|
| 1178 |
"train_batch_size": 128,
|
| 1179 |
"trial_name": null,
|
| 1180 |
"trial_params": null
|