Training in progress, step 5000, checkpoint

last-checkpoint/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9bd26932b7a7d214154cb540340ba893a284b77c131735da01b6ac665eb47be6
+oid sha256:776e810dd01ff44cd0ae5258784cc7a4fb43912af9e20ebd4fb14a16e6edf73a
 size 409608164

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e5f340db329937064109427304f9c4f25af0287f4271f0d77c733d5a47e70637
+oid sha256:52449a96f9d7515622a053360c3cb75fff069929f11d88163ad9b3f90c1e0a21
 size 814647162

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:466f74c126878bc66457d9c7e1b729fc78dc8be9b74baf7965800ea92bbcba9b
+oid sha256:43f004ece9bfea3a082d65f9d4e84c653ebe3edebf4e07157b3b5a118e85e9b9
 size 14244

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:44bf41f769bbe8aab9b52c61c05b6230543dfc2e7fd8b060d0f2955e2430ab77
+oid sha256:77441d8a72ed30d9525730698a7169f1229749dee2b43dbae89633553d79ba30
 size 1064

last-checkpoint/trainer_state.json

@@ -1,278 +1,124 @@
 {
-  "best_metric": 0.0,
-  "best_model_checkpoint": "/mimer/NOBACKUP/groups/naiss2023-6-290/stefano/models//PROTAC-Splitter-EncoderDecoder-lr_cosine-opt25/trial-number=23-learning_rate=5.9e-04-warmup_ratio=0.070-num_cycles=4.500/checkpoint-5000",
-  "epoch": 0.9863878477017163,
+  "best_metric": 0.003977021652673442,
+  "best_model_checkpoint": "/mimer/NOBACKUP/groups/naiss2023-6-290/stefano/models//PROTAC-Splitter-EncoderDecoder-lr_cosine-opt25/trial-number=24-learning_rate=1.2e-05-warmup_ratio=0.020-num_cycles=8.500/checkpoint-5000",
+  "epoch": 0.49319392385085814,
   "eval_steps": 2500,
-  "global_step": 10000,
+  "global_step": 5000,
   "is_hyper_param_search": true,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.049319392385085814,
-      "grad_norm": 0.489678293466568,
-      "learning_rate": 0.00042135820992193263,
-      "loss": 1.109,
+      "grad_norm": 1.171087622642517,
+      "learning_rate": 8.701946418116152e-06,
+      "loss": 3.6793,
       "step": 500
     },
     {
       "epoch": 0.09863878477017163,
-      "grad_norm": 0.1939166784286499,
-      "learning_rate": 0.0004741266311129676,
-      "loss": 0.0641,
+      "grad_norm": 1.163271188735962,
+      "learning_rate": 5.128180511749551e-06,
+      "loss": 0.8322,
       "step": 1000
     },
     {
       "epoch": 0.14795817715525744,
-      "grad_norm": 0.09205025434494019,
-      "learning_rate": 7.052064965899953e-05,
-      "loss": 0.0156,
+      "grad_norm": 1.0759015083312988,
+      "learning_rate": 5.396879384932859e-06,
+      "loss": 0.5681,
       "step": 1500
     },
     {
       "epoch": 0.19727756954034326,
-      "grad_norm": 0.06918811798095703,
-      "learning_rate": 9.216730521484801e-05,
-      "loss": 0.0067,
+      "grad_norm": 0.8901547193527222,
+      "learning_rate": 8.358809829778745e-06,
+      "loss": 0.4439,
       "step": 2000
     },
     {
       "epoch": 0.24659696192542907,
-      "grad_norm": 0.4081190228462219,
-      "learning_rate": 0.0004979731235012487,
-      "loss": 0.0224,
+      "grad_norm": 0.6830881834030151,
+      "learning_rate": 2.3817412792462764e-06,
+      "loss": 0.3258,
       "step": 2500
     },
     {
       "epoch": 0.2959163543105149,
-      "grad_norm": 0.6822362542152405,
-      "learning_rate": 0.0005175663991476457,
-      "loss": 0.1633,
+      "grad_norm": 0.8356934785842896,
+      "learning_rate": 1.0939094459938762e-05,
+      "loss": 0.285,
       "step": 3000
     },
     {
       "epoch": 0.3452357466956007,
-      "grad_norm": 0.3597358763217926,
-      "learning_rate": 0.00011375174339089069,
-      "loss": 0.1024,
+      "grad_norm": 0.437906414270401,
+      "learning_rate": 4.240352158990673e-07,
+      "loss": 0.2182,
       "step": 3500
     },
     {
       "epoch": 0.3945551390806865,
-      "grad_norm": 0.37279772758483887,
-      "learning_rate": 5.312088801685656e-05,
-      "loss": 0.0736,
+      "grad_norm": 0.5716018080711365,
+      "learning_rate": 1.2131791720228002e-05,
+      "loss": 0.2003,
       "step": 4000
     },
     {
       "epoch": 0.4438745314657723,
-      "grad_norm": 0.34280383586883545,
-      "learning_rate": 0.00045077394580807,
-      "loss": 0.0759,
+      "grad_norm": 0.6657638549804688,
+      "learning_rate": 8.311982216924161e-08,
+      "loss": 0.1613,
       "step": 4500
     },
     {
       "epoch": 0.49319392385085814,
-      "grad_norm": 0.26311400532722473,
-      "learning_rate": 0.0005518162098083314,
-      "loss": 0.1777,
+      "grad_norm": 0.6562560796737671,
+      "learning_rate": 1.159612241260229e-05,
+      "loss": 0.1497,
       "step": 5000
     },
     {
       "epoch": 0.49319392385085814,
-      "eval_all_ligands_equal": 0.0,
-      "eval_e3_equal": 0.015377817057003976,
+      "eval_all_ligands_equal": 0.003977021652673442,
+      "eval_e3_equal": 0.38868758285461774,
       "eval_e3_graph_edit_distance": 9.999999999999999e+63,
       "eval_e3_graph_edit_distance_norm": 1.0,
-      "eval_e3_has_attachment_point(s)": 1.0,
-      "eval_e3_heavy_atoms_difference": 9.350861688024747,
-      "eval_e3_heavy_atoms_difference_norm": 0.2672570344367313,
+      "eval_e3_has_attachment_point(s)": 0.9946973044631021,
+      "eval_e3_heavy_atoms_difference": -0.6082191780821918,
+      "eval_e3_heavy_atoms_difference_norm": -0.039596567096754735,
       "eval_e3_tanimoto_similarity": 0.0,
-      "eval_e3_valid": 1.0,
-      "eval_has_all_attachment_points": 1.0,
-      "eval_has_three_substructures": 1.0,
-      "eval_heavy_atoms_difference": 14.4830755634114,
-      "eval_heavy_atoms_difference_norm": 0.15956557502921162,
-      "eval_linker_equal": 0.0009721608484312859,
+      "eval_e3_valid": 0.9946973044631021,
+      "eval_has_all_attachment_points": 0.9510384445426425,
+      "eval_has_three_substructures": 0.9971718957136544,
+      "eval_heavy_atoms_difference": 19.936544410075122,
+      "eval_heavy_atoms_difference_norm": 0.2585121211609394,
+      "eval_linker_equal": 0.018471056120194434,
       "eval_linker_graph_edit_distance": 9.999999999999999e+63,
       "eval_linker_graph_edit_distance_norm": 1.0,
-      "eval_linker_has_attachment_point(s)": 1.0,
-      "eval_linker_heavy_atoms_difference": -0.38700839593460007,
-      "eval_linker_heavy_atoms_difference_norm": -0.2639517784176107,
+      "eval_linker_has_attachment_point(s)": 0.9883340698188245,
+      "eval_linker_heavy_atoms_difference": 5.180998674326116,
+      "eval_linker_heavy_atoms_difference_norm": 0.23542542857620521,
       "eval_linker_tanimoto_similarity": 0.0,
-      "eval_linker_valid": 1.0,
-      "eval_loss": 0.5065851807594299,
-      "eval_num_fragments": 3.0,
-      "eval_poi_equal": 0.002209456473707468,
+      "eval_linker_valid": 0.9883340698188245,
+      "eval_loss": 0.3508746027946472,
+      "eval_num_fragments": 2.998320813079982,
+      "eval_poi_equal": 0.35899248784798937,
       "eval_poi_graph_edit_distance": 9.999999999999999e+63,
       "eval_poi_graph_edit_distance_norm": 1.0,
-      "eval_poi_has_attachment_point(s)": 1.0,
-      "eval_poi_heavy_atoms_difference": 5.519222271321255,
-      "eval_poi_heavy_atoms_difference_norm": 0.05816090379761925,
+      "eval_poi_has_attachment_point(s)": 0.8076005302695537,
+      "eval_poi_heavy_atoms_difference": 6.755987627043747,
+      "eval_poi_heavy_atoms_difference_norm": 0.2056052919584098,
       "eval_poi_tanimoto_similarity": 0.0,
-      "eval_poi_valid": 1.0,
-      "eval_reassembly": 0.0,
-      "eval_reassembly_nostereo": 0.0,
-      "eval_runtime": 1049.5929,
-      "eval_samples_per_second": 10.78,
-      "eval_steps_per_second": 0.169,
+      "eval_poi_valid": 0.8076005302695537,
+      "eval_reassembly": 0.004065399911621741,
+      "eval_reassembly_nostereo": 0.005832965090587715,
+      "eval_runtime": 1511.2417,
+      "eval_samples_per_second": 7.487,
+      "eval_steps_per_second": 0.117,
       "eval_tanimoto_similarity": 0.0,
-      "eval_valid": 1.0,
+      "eval_valid": 0.8003535130357932,
       "step": 5000
-    },
-    {
-      "epoch": 0.542513316235944,
-      "grad_norm": 0.21691560745239258,
-      "learning_rate": 0.00016443155091165759,
-      "loss": 0.163,
-      "step": 5500
-    },
-    {
-      "epoch": 0.5918327086210298,
-      "grad_norm": 0.3181842267513275,
-      "learning_rate": 2.4021402169957885e-05,
-      "loss": 0.1513,
-      "step": 6000
-    },
-    {
-      "epoch": 0.6411521010061156,
-      "grad_norm": 0.17199087142944336,
-      "learning_rate": 0.0003971369091390999,
-      "loss": 0.1536,
-      "step": 6500
-    },
-    {
-      "epoch": 0.6904714933912014,
-      "grad_norm": 0.29000797867774963,
-      "learning_rate": 0.000575732416024195,
-      "loss": 0.2037,
-      "step": 7000
-    },
-    {
-      "epoch": 0.7397908857762873,
-      "grad_norm": 0.211964949965477,
-      "learning_rate": 0.00022133879707214477,
-      "loss": 0.1599,
-      "step": 7500
-    },
-    {
-      "epoch": 0.7397908857762873,
-      "eval_all_ligands_equal": 0.0,
-      "eval_e3_equal": 0.007512152010605391,
-      "eval_e3_graph_edit_distance": 9.999999999999999e+63,
-      "eval_e3_graph_edit_distance_norm": 1.0,
-      "eval_e3_has_attachment_point(s)": 1.0,
-      "eval_e3_heavy_atoms_difference": -5.001325673884224,
-      "eval_e3_heavy_atoms_difference_norm": -0.29136050626839866,
-      "eval_e3_tanimoto_similarity": 0.0,
-      "eval_e3_valid": 1.0,
-      "eval_has_all_attachment_points": 1.0,
-      "eval_has_three_substructures": 1.0,
-      "eval_heavy_atoms_difference": -15.595139195757843,
-      "eval_heavy_atoms_difference_norm": -0.25304102034396836,
-      "eval_linker_equal": 0.0013256738842244808,
-      "eval_linker_graph_edit_distance": 9.999999999999999e+63,
-      "eval_linker_graph_edit_distance_norm": 1.0,
-      "eval_linker_has_attachment_point(s)": 1.0,
-      "eval_linker_heavy_atoms_difference": -2.37870083959346,
-      "eval_linker_heavy_atoms_difference_norm": -0.41310975884925333,
-      "eval_linker_tanimoto_similarity": 0.0,
-      "eval_linker_valid": 1.0,
-      "eval_loss": 0.5110778212547302,
-      "eval_num_fragments": 3.0,
-      "eval_poi_equal": 0.0017675651789659744,
-      "eval_poi_graph_edit_distance": 9.999999999999999e+63,
-      "eval_poi_graph_edit_distance_norm": 1.0,
-      "eval_poi_has_attachment_point(s)": 1.0,
-      "eval_poi_heavy_atoms_difference": -8.215112682280159,
-      "eval_poi_heavy_atoms_difference_norm": -0.4444324339014155,
-      "eval_poi_tanimoto_similarity": 0.0,
-      "eval_poi_valid": 1.0,
-      "eval_reassembly": 0.0,
-      "eval_reassembly_nostereo": 0.0,
-      "eval_runtime": 1528.456,
-      "eval_samples_per_second": 7.403,
-      "eval_steps_per_second": 0.116,
-      "eval_tanimoto_similarity": 0.0,
-      "eval_valid": 1.0,
-      "step": 7500
-    },
-    {
-      "epoch": 0.789110278161373,
-      "grad_norm": 0.24591299891471863,
-      "learning_rate": 5.888382913999639e-06,
-      "loss": 0.1498,
-      "step": 8000
-    },
-    {
-      "epoch": 0.8384296705464589,
-      "grad_norm": 0.22677373886108398,
-      "learning_rate": 0.0003383868986984896,
-      "loss": 0.1503,
-      "step": 8500
-    },
-    {
-      "epoch": 0.8877490629315447,
-      "grad_norm": 0.4033995568752289,
-      "learning_rate": 0.0005876274048117072,
-      "loss": 0.1776,
-      "step": 9000
-    },
-    {
-      "epoch": 0.9370684553166305,
-      "grad_norm": 0.22147516906261444,
-      "learning_rate": 0.0002804579032052931,
-      "loss": 0.1634,
-      "step": 9500
-    },
-    {
-      "epoch": 0.9863878477017163,
-      "grad_norm": 0.24389365315437317,
-      "learning_rate": 1.3608923772205227e-09,
-      "loss": 0.149,
-      "step": 10000
-    },
-    {
-      "epoch": 0.9863878477017163,
-      "eval_all_ligands_equal": 0.0,
-      "eval_e3_equal": 0.006098099867432612,
-      "eval_e3_graph_edit_distance": 9.999999999999999e+63,
-      "eval_e3_graph_edit_distance_norm": 1.0,
-      "eval_e3_has_attachment_point(s)": 1.0,
-      "eval_e3_heavy_atoms_difference": -7.204772425983208,
-      "eval_e3_heavy_atoms_difference_norm": -0.3807529197617233,
-      "eval_e3_tanimoto_similarity": 0.0,
-      "eval_e3_valid": 1.0,
-      "eval_has_all_attachment_points": 1.0,
-      "eval_has_three_substructures": 1.0,
-      "eval_heavy_atoms_difference": -25.552452496685817,
-      "eval_heavy_atoms_difference_norm": -0.39024082574687935,
-      "eval_linker_equal": 0.0016791869200176756,
-      "eval_linker_graph_edit_distance": 9.999999999999999e+63,
-      "eval_linker_graph_edit_distance_norm": 1.0,
-      "eval_linker_has_attachment_point(s)": 1.0,
-      "eval_linker_heavy_atoms_difference": -3.351568714096332,
-      "eval_linker_heavy_atoms_difference_norm": -0.4874536631163987,
-      "eval_linker_tanimoto_similarity": 0.0,
-      "eval_linker_valid": 1.0,
-      "eval_loss": 0.5194833278656006,
-      "eval_num_fragments": 3.0,
-      "eval_poi_equal": 0.0013256738842244808,
-      "eval_poi_graph_edit_distance": 9.999999999999999e+63,
-      "eval_poi_graph_edit_distance_norm": 1.0,
-      "eval_poi_has_attachment_point(s)": 1.0,
-      "eval_poi_heavy_atoms_difference": -14.996111356606274,
-      "eval_poi_heavy_atoms_difference_norm": -0.6893647255577231,
-      "eval_poi_tanimoto_similarity": 0.0,
-      "eval_poi_valid": 1.0,
-      "eval_reassembly": 0.0,
-      "eval_reassembly_nostereo": 0.0,
-      "eval_runtime": 1742.4478,
-      "eval_samples_per_second": 6.494,
-      "eval_steps_per_second": 0.102,
-      "eval_tanimoto_similarity": 0.0,
-      "eval_valid": 1.0,
-      "step": 10000
     }
   ],
   "logging_steps": 500,
@@ -287,17 +133,17 @@
         "should_evaluate": false,
         "should_log": false,
         "should_save": true,
-        "should_training_stop": true
+        "should_training_stop": false
       },
       "attributes": {}
     }
   },
-  "total_flos": 8.663036348367667e+16,
+  "total_flos": 4.33139477262336e+16,
   "train_batch_size": 128,
-  "trial_name": "trial-number=23-learning_rate=5.9e-04-warmup_ratio=0.070-num_cycles=4.500",
+  "trial_name": "trial-number=24-learning_rate=1.2e-05-warmup_ratio=0.020-num_cycles=8.500",
   "trial_params": {
-    "learning_rate": 0.0005890587774708618,
-    "num_cycles": 4.5,
-    "warmup_ratio": 0.06999999999999999
+    "learning_rate": 1.2165321092526382e-05,
+    "num_cycles": 8.5,
+    "warmup_ratio": 0.02
   }
 }

last-checkpoint/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a52203b3423a00dd886f1d1e0cb64df3ed566569148065df4628b517d3970520
+oid sha256:2019bae8376c62a133e2b8843d20a676fb423a0cdd62225bc692109b5e9ebb28
 size 7416