Training in progress, step 10, checkpoint

last-checkpoint/adapter_config.json

@@ -20,13 +20,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "down_proj",
-    "up_proj",
-    "k_proj",
     "v_proj",
-    "o_proj",
+    "k_proj",
+    "q_proj",
+    "up_proj",
     "gate_proj",
-    "q_proj"
+    "down_proj",
+    "o_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

last-checkpoint/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a5eb5aa49deee1c7b9ed99237d89de35a36f9bb595761637f2bf81c7c5db0b96
+oid sha256:8d1651f3601378af486d8705149df9a51b2def80ca66d31c03b802020e5a8db9
 size 83115256

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8cb3c6399204c6987be382b0054ab53cc0a980d62814f30caa7b35502d5f4df3
+oid sha256:0b50128d3b0c6a2edc34f6c7aabbd7e35a705b84c33a17a46e29421d95323100
 size 166351098

last-checkpoint/trainer_state.json

@@ -10,7 +10,7 @@
   "log_history": [
     {
       "epoch": 0.0005925925925925926,
-      "grad_norm": 10.858694076538086,
+      "grad_norm": 10.913461685180664,
       "learning_rate": 5e-07,
       "logits/chosen": -4.444676399230957,
       "logits/rejected": -4.0909342765808105,
@@ -25,7 +25,7 @@
     },
     {
       "epoch": 0.0011851851851851852,
-      "grad_norm": 7.565941333770752,
+      "grad_norm": 7.564138412475586,
       "learning_rate": 4.849231551964771e-07,
       "logits/chosen": -4.530362606048584,
       "logits/rejected": -3.981240749359131,
@@ -40,122 +40,122 @@
     },
     {
       "epoch": 0.0017777777777777779,
-      "grad_norm": 13.622398376464844,
+      "grad_norm": 14.393203735351562,
       "learning_rate": 4.415111107797445e-07,
-      "logits/chosen": -4.5415449142456055,
-      "logits/rejected": -4.322567462921143,
-      "logps/chosen": -200.85960388183594,
-      "logps/rejected": -224.73626708984375,
-      "loss": 0.6619,
+      "logits/chosen": -4.539660930633545,
+      "logits/rejected": -4.32346248626709,
+      "logps/chosen": -200.69912719726562,
+      "logps/rejected": -224.77163696289062,
+      "loss": 0.6526,
       "rewards/accuracies": 1.0,
-      "rewards/chosen": 0.03391075134277344,
-      "rewards/margins": 0.06416893005371094,
-      "rewards/rejected": -0.03025818057358265,
+      "rewards/chosen": 0.04995880275964737,
+      "rewards/margins": 0.08375511318445206,
+      "rewards/rejected": -0.03379631042480469,
       "step": 3
     },
     {
       "epoch": 0.0023703703703703703,
-      "grad_norm": 10.704012870788574,
+      "grad_norm": 12.62684440612793,
       "learning_rate": 3.75e-07,
-      "logits/chosen": -3.9766385555267334,
-      "logits/rejected": -4.474390029907227,
-      "logps/chosen": -239.95339965820312,
-      "logps/rejected": -202.6293182373047,
-      "loss": 0.7176,
-      "rewards/accuracies": 0.25,
-      "rewards/chosen": -0.037302400916814804,
-      "rewards/margins": -0.04801139980554581,
-      "rewards/rejected": 0.010708999820053577,
+      "logits/chosen": -3.9760360717773438,
+      "logits/rejected": -4.4724321365356445,
+      "logps/chosen": -239.9991455078125,
+      "logps/rejected": -202.63072204589844,
+      "loss": 0.7198,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": -0.041876986622810364,
+      "rewards/margins": -0.052445217967033386,
+      "rewards/rejected": 0.01056823693215847,
       "step": 4
     },
     {
       "epoch": 0.002962962962962963,
-      "grad_norm": 6.752383232116699,
+      "grad_norm": 6.745312213897705,
       "learning_rate": 2.934120444167326e-07,
-      "logits/chosen": -4.133105754852295,
-      "logits/rejected": -4.693289279937744,
-      "logps/chosen": -168.95433044433594,
-      "logps/rejected": -149.30447387695312,
-      "loss": 0.6772,
-      "rewards/accuracies": 0.75,
-      "rewards/chosen": 0.03451080620288849,
-      "rewards/margins": 0.03242149576544762,
-      "rewards/rejected": 0.002089309971779585,
+      "logits/chosen": -4.136646270751953,
+      "logits/rejected": -4.701557159423828,
+      "logps/chosen": -169.00192260742188,
+      "logps/rejected": -149.2642059326172,
+      "loss": 0.6818,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": 0.02975158765912056,
+      "rewards/margins": 0.023634720593690872,
+      "rewards/rejected": 0.006116867531090975,
       "step": 5
     },
     {
       "epoch": 0.0035555555555555557,
-      "grad_norm": 12.878663063049316,
+      "grad_norm": 13.741002082824707,
       "learning_rate": 2.065879555832674e-07,
-      "logits/chosen": -5.0531182289123535,
-      "logits/rejected": -5.144861221313477,
-      "logps/chosen": -193.86325073242188,
-      "logps/rejected": -202.404052734375,
-      "loss": 0.6904,
-      "rewards/accuracies": 0.5,
-      "rewards/chosen": -0.009803391061723232,
-      "rewards/margins": 0.0061431871727108955,
-      "rewards/rejected": -0.015946578234434128,
+      "logits/chosen": -5.057063579559326,
+      "logits/rejected": -5.147495746612549,
+      "logps/chosen": -193.6820526123047,
+      "logps/rejected": -202.20425415039062,
+      "loss": 0.6913,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": 0.0083160400390625,
+      "rewards/margins": 0.004281995818018913,
+      "rewards/rejected": 0.0040340423583984375,
       "step": 6
     },
     {
       "epoch": 0.004148148148148148,
-      "grad_norm": 6.771181106567383,
+      "grad_norm": 6.920108795166016,
       "learning_rate": 1.2500000000000005e-07,
-      "logits/chosen": -4.2494401931762695,
-      "logits/rejected": -3.684239387512207,
-      "logps/chosen": -176.15640258789062,
-      "logps/rejected": -223.1680450439453,
-      "loss": 0.6937,
-      "rewards/accuracies": 0.25,
-      "rewards/chosen": 0.02459411695599556,
-      "rewards/margins": -0.0004608146846294403,
-      "rewards/rejected": 0.025054931640625,
+      "logits/chosen": -4.2448506355285645,
+      "logits/rejected": -3.682173490524292,
+      "logps/chosen": -176.11734008789062,
+      "logps/rejected": -222.982666015625,
+      "loss": 0.7008,
+      "rewards/accuracies": 0.0,
+      "rewards/chosen": 0.028499603271484375,
+      "rewards/margins": -0.015093998052179813,
+      "rewards/rejected": 0.043593600392341614,
       "step": 7
     },
     {
       "epoch": 0.004740740740740741,
-      "grad_norm": 11.16915225982666,
+      "grad_norm": 10.856575965881348,
       "learning_rate": 5.848888922025552e-08,
-      "logits/chosen": -4.407987117767334,
-      "logits/rejected": -4.415105819702148,
-      "logps/chosen": -182.2259063720703,
-      "logps/rejected": -196.61068725585938,
-      "loss": 0.6884,
-      "rewards/accuracies": 0.5,
-      "rewards/chosen": -0.05178241804242134,
-      "rewards/margins": 0.010274123400449753,
-      "rewards/rejected": -0.062056541442871094,
+      "logits/chosen": -4.413946151733398,
+      "logits/rejected": -4.419940948486328,
+      "logps/chosen": -181.99610900878906,
+      "logps/rejected": -196.20811462402344,
+      "loss": 0.6967,
+      "rewards/accuracies": 0.25,
+      "rewards/chosen": -0.02880115620791912,
+      "rewards/margins": -0.007002831436693668,
+      "rewards/rejected": -0.021798323839902878,
       "step": 8
     },
     {
       "epoch": 0.005333333333333333,
-      "grad_norm": 9.21722412109375,
+      "grad_norm": 9.559263229370117,
       "learning_rate": 1.507684480352292e-08,
-      "logits/chosen": -4.9174699783325195,
-      "logits/rejected": -4.600230693817139,
-      "logps/chosen": -154.47506713867188,
-      "logps/rejected": -178.89224243164062,
-      "loss": 0.7014,
+      "logits/chosen": -4.909343242645264,
+      "logits/rejected": -4.601991176605225,
+      "logps/chosen": -154.6862030029297,
+      "logps/rejected": -179.27401733398438,
+      "loss": 0.6929,
       "rewards/accuracies": 0.25,
-      "rewards/chosen": 0.024329185485839844,
-      "rewards/margins": -0.01587962917983532,
-      "rewards/rejected": 0.04020881652832031,
+      "rewards/chosen": 0.0032157916575670242,
+      "rewards/margins": 0.0011837054044008255,
+      "rewards/rejected": 0.0020320871844887733,
       "step": 9
     },
     {
       "epoch": 0.005925925925925926,
-      "grad_norm": 16.037979125976562,
+      "grad_norm": 10.816996574401855,
       "learning_rate": 0.0,
-      "logits/chosen": -3.811877965927124,
-      "logits/rejected": -3.696570873260498,
-      "logps/chosen": -232.4647216796875,
-      "logps/rejected": -253.81515502929688,
-      "loss": 0.7093,
-      "rewards/accuracies": 0.25,
-      "rewards/chosen": -0.00266952533274889,
-      "rewards/margins": -0.03174018859863281,
-      "rewards/rejected": 0.029070664197206497,
+      "logits/chosen": -3.8037424087524414,
+      "logits/rejected": -3.683922529220581,
+      "logps/chosen": -232.978515625,
+      "logps/rejected": -254.24078369140625,
+      "loss": 0.7146,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.05405044183135033,
+      "rewards/margins": -0.040559008717536926,
+      "rewards/rejected": -0.013491439633071423,
       "step": 10
     }
   ],

last-checkpoint/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3a8e9b719f5f1bef02ca8d4f429a8f8a2f1e58ec49b254b8eeb60d43f830af0f
+oid sha256:500a27bb31e08eab12d9fb299fbdfd1af99d44a30059f736748d9f929e4cbfa9
 size 5944