Training in progress, step 70000, checkpoint

last-checkpoint/generation_config.json

@@ -1,5 +1,4 @@
 {
-  "_from_model_config": true,
   "bos_token_id": 0,
   "do_sample": true,
   "eos_token_id": 2,

last-checkpoint/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b765806c32a71debbe5fcd75a03b9d683e507d65faff94e01c59c7581ed7ac5f
+oid sha256:81fabb91cc7e3f0e7e2d6a6b39656d9b97dcf407d5e66728f3eace40a6625685
 size 409608164

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:973086399ebf7dfee4ca3fae615f7dbbcc6764ef896837b42c878ef225b08ff0
+oid sha256:0708124113d024d17d7e8725c69fa04f56a7afc2eb33b6474a8454cdce7e30f5
 size 814647162

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d94c3d1ced5d6eccb982de07377a5bac97c708bdc2a2e2ec8462b5f4c3224ec7
+oid sha256:2843dadbca00c0a408cf867f0abe54c839025990f79836b4c579a8f5012d3914
 size 14244

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1ef1f0d7258e92b9f9871d81d52ebcc946c37081ef596ae367fa8c041eca7a05
+oid sha256:d7b405293c7c61627577f0ffca1ebd7a6cc00745b4ef6a4df74ea0a9189fd0ef
 size 1064

last-checkpoint/trainer_state.json

@@ -1,9 +1,9 @@
 {
-  "best_metric": 0.5389518413597734,
-  "best_model_checkpoint": "/mimer/NOBACKUP/groups/naiss2023-6-290/stefano/models//PROTAC-Splitter-EncoderDecoder-lr_cosine-opt25-rand-smiles/checkpoint-10000",
-  "epoch": 6.411521010061156,
+  "best_metric": 0.5924220963172805,
+  "best_model_checkpoint": "/mimer/NOBACKUP/groups/naiss2023-6-290/stefano/models//PROTAC-Splitter-EncoderDecoder-lr_cosine-rand-smiles/checkpoint-70000",
+  "epoch": 6.904714933912015,
   "eval_steps": 2500,
-  "global_step": 65000,
+  "global_step": 70000,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -1043,6 +1043,118 @@
       "learning_rate": 1.3684575241351546e-05,
       "loss": 0.0004,
       "step": 65000
+    },
+    {
+      "epoch": 6.460840402446242,
+      "grad_norm": 0.02038782648742199,
+      "learning_rate": 1.3335466168024479e-05,
+      "loss": 0.0004,
+      "step": 65500
+    },
+    {
+      "epoch": 6.510159794831328,
+      "grad_norm": 0.005516901146620512,
+      "learning_rate": 1.2989240908505564e-05,
+      "loss": 0.0004,
+      "step": 66000
+    },
+    {
+      "epoch": 6.5594791872164135,
+      "grad_norm": 0.0024911893997341394,
+      "learning_rate": 1.2645985059801791e-05,
+      "loss": 0.0004,
+      "step": 66500
+    },
+    {
+      "epoch": 6.608798579601499,
+      "grad_norm": 0.021164007484912872,
+      "learning_rate": 1.2306460783748302e-05,
+      "loss": 0.0004,
+      "step": 67000
+    },
+    {
+      "epoch": 6.658117971986585,
+      "grad_norm": 0.029552442952990532,
+      "learning_rate": 1.196939123002433e-05,
+      "loss": 0.0004,
+      "step": 67500
+    },
+    {
+      "epoch": 6.707437364371671,
+      "grad_norm": 0.010598881170153618,
+      "learning_rate": 1.1635543223774156e-05,
+      "loss": 0.0004,
+      "step": 68000
+    },
+    {
+      "epoch": 6.756756756756757,
+      "grad_norm": 0.00585036538541317,
+      "learning_rate": 1.130499930198588e-05,
+      "loss": 0.0004,
+      "step": 68500
+    },
+    {
+      "epoch": 6.806076149141843,
+      "grad_norm": 0.011820879764854908,
+      "learning_rate": 1.0977841184781343e-05,
+      "loss": 0.0003,
+      "step": 69000
+    },
+    {
+      "epoch": 6.855395541526929,
+      "grad_norm": 0.03137628361582756,
+      "learning_rate": 1.0654793624981977e-05,
+      "loss": 0.0003,
+      "step": 69500
+    },
+    {
+      "epoch": 6.904714933912015,
+      "grad_norm": 0.013417651876807213,
+      "learning_rate": 1.0334641736320572e-05,
+      "loss": 0.0003,
+      "step": 70000
+    },
+    {
+      "epoch": 6.904714933912015,
+      "eval_all_ligands_equal": 0.5924220963172805,
+      "eval_e3_equal": 0.8257790368271954,
+      "eval_e3_graph_edit_distance": Infinity,
+      "eval_e3_graph_edit_distance_norm": Infinity,
+      "eval_e3_has_attachment_point(s)": 0.9915014164305949,
+      "eval_e3_heavy_atoms_difference": 0.14606940509915015,
+      "eval_e3_heavy_atoms_difference_norm": -0.002032805775238171,
+      "eval_e3_tanimoto_similarity": 0.0,
+      "eval_e3_valid": 0.9915014164305949,
+      "eval_has_all_attachment_points": 0.9909702549575071,
+      "eval_has_three_substructures": 0.9996458923512748,
+      "eval_heavy_atoms_difference": 4.543112606232294,
+      "eval_heavy_atoms_difference_norm": 0.06065633437693917,
+      "eval_linker_equal": 0.8415368271954674,
+      "eval_linker_graph_edit_distance": 2.2131728045325776e+61,
+      "eval_linker_graph_edit_distance_norm": Infinity,
+      "eval_linker_has_attachment_point(s)": 0.9977868271954674,
+      "eval_linker_heavy_atoms_difference": 0.16979461756373937,
+      "eval_linker_heavy_atoms_difference_norm": -0.0016392134755984211,
+      "eval_linker_tanimoto_similarity": 0.0,
+      "eval_linker_valid": 0.9977868271954674,
+      "eval_loss": 0.39393511414527893,
+      "eval_num_fragments": 3.0003541076487252,
+      "eval_poi_equal": 0.7886862606232294,
+      "eval_poi_graph_edit_distance": Infinity,
+      "eval_poi_graph_edit_distance_norm": Infinity,
+      "eval_poi_has_attachment_point(s)": 0.9554709631728046,
+      "eval_poi_heavy_atoms_difference": 1.2542492917847026,
+      "eval_poi_heavy_atoms_difference_norm": 0.034870780705027825,
+      "eval_poi_tanimoto_similarity": 0.0,
+      "eval_poi_valid": 0.9554709631728046,
+      "eval_reassembly": 0.5998583569405099,
+      "eval_reassembly_nostereo": 0.6323477337110481,
+      "eval_runtime": 2303.8454,
+      "eval_samples_per_second": 4.903,
+      "eval_steps_per_second": 0.077,
+      "eval_tanimoto_similarity": 0.0,
+      "eval_valid": 0.9464412181303116,
+      "step": 70000
     }
   ],
   "logging_steps": 500,
@@ -1062,7 +1174,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 5.674161087565732e+17,
+  "total_flos": 6.110533685712476e+17,
   "train_batch_size": 128,
   "trial_name": null,
   "trial_params": null