Training in progress, step 750, checkpoint

last-checkpoint/adapter_config.json

@@ -22,8 +22,8 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "q_proj",
-    "k_proj"
+    "k_proj",
+    "q_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 sha256:593eb37c50a4439ba03e83ada103a9e73c7a6cc5e2190f9a48dcfe15e287b91a
+oid sha256:ca2c73898824981f7e0729fc57be3e758b804c6877cb01d8db2421cd68ab9ec7
 size 527048968

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:540ae6c9df4b9fc1733ab8364067bef7b8eed491e5133d225a8fa424852ac127
+oid sha256:bfa421196ea17ada7bbfeeaa47740f883958c13c1c7bf744610376f3f45e10dc
 size 1054135994

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:280ae9c0809eceeb464f568a46cee1cc6c997124b85d7c1db2a216ff1d3932aa
+oid sha256:30b57ec052aeaf035e39cd5057ce6977dfdadd2be87b9e8aaa1d9afd55d2f8e1
 size 14244

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b5ff073b2d9654300593116ff77249d188bd5752ecbe3d0cb800964ea5d21c80
+oid sha256:e8974fc0327a603be8a90ee235d3ef81c4ee4af2818c873bdeb9bfb38f108e1a
 size 1064

last-checkpoint/trainer_state.json

@@ -1,464 +1,577 @@
 {
-  "best_metric": 0.7168648838996887,
+  "best_metric": 0.7164609432220459,
   "best_model_checkpoint": "./output/checkpoint-450",
-  "epoch": 26.08695652173913,
+  "epoch": 32.608695652173914,
   "eval_steps": 150,
-  "global_step": 600,
+  "global_step": 750,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.43478260869565216,
-      "grad_norm": 1.5021412372589111,
+      "grad_norm": 1.5021647214889526,
       "learning_rate": 3e-06,
-      "loss": 0.9061,
+      "loss": 0.9064,
       "step": 10
     },
     {
       "epoch": 0.8695652173913043,
-      "grad_norm": 1.687071442604065,
+      "grad_norm": 1.6793277263641357,
       "learning_rate": 6e-06,
-      "loss": 0.9026,
+      "loss": 0.9024,
       "step": 20
     },
     {
       "epoch": 1.3043478260869565,
-      "grad_norm": 1.7296651601791382,
+      "grad_norm": 1.7296056747436523,
       "learning_rate": 9e-06,
-      "loss": 0.9003,
+      "loss": 0.9,
       "step": 30
     },
     {
       "epoch": 1.7391304347826086,
-      "grad_norm": 1.4536513090133667,
+      "grad_norm": 1.4458602666854858,
       "learning_rate": 1.2e-05,
-      "loss": 0.9093,
+      "loss": 0.9086,
       "step": 40
     },
     {
       "epoch": 2.1739130434782608,
-      "grad_norm": 1.351776361465454,
+      "grad_norm": 1.3516699075698853,
       "learning_rate": 1.5e-05,
-      "loss": 0.8358,
+      "loss": 0.8359,
       "step": 50
     },
     {
       "epoch": 2.608695652173913,
-      "grad_norm": 2.031055212020874,
+      "grad_norm": 2.03114914894104,
       "learning_rate": 1.8e-05,
-      "loss": 0.8894,
+      "loss": 0.8892,
       "step": 60
     },
     {
       "epoch": 3.0434782608695654,
-      "grad_norm": 1.4920436143875122,
+      "grad_norm": 1.4920729398727417,
       "learning_rate": 2.1e-05,
-      "loss": 0.8915,
+      "loss": 0.8913,
       "step": 70
     },
     {
       "epoch": 3.4782608695652173,
-      "grad_norm": 1.729347586631775,
+      "grad_norm": 1.7293280363082886,
       "learning_rate": 2.4e-05,
-      "loss": 0.8237,
+      "loss": 0.8234,
       "step": 80
     },
     {
       "epoch": 3.9130434782608696,
-      "grad_norm": 1.4314346313476562,
+      "grad_norm": 1.4240705966949463,
       "learning_rate": 2.7000000000000002e-05,
       "loss": 0.8529,
       "step": 90
     },
     {
       "epoch": 4.3478260869565215,
-      "grad_norm": 1.3657989501953125,
+      "grad_norm": 1.3656291961669922,
       "learning_rate": 3e-05,
       "loss": 0.8649,
       "step": 100
     },
     {
       "epoch": 4.782608695652174,
-      "grad_norm": 2.1989877223968506,
+      "grad_norm": 2.198136806488037,
       "learning_rate": 2.999999702723963e-05,
       "loss": 0.8227,
       "step": 110
     },
     {
       "epoch": 5.217391304347826,
-      "grad_norm": 1.0724585056304932,
+      "grad_norm": 1.0726988315582275,
       "learning_rate": 2.9999988108959687e-05,
-      "loss": 0.7655,
+      "loss": 0.765,
       "step": 120
     },
     {
       "epoch": 5.6521739130434785,
-      "grad_norm": 1.5675039291381836,
+      "grad_norm": 1.5672078132629395,
       "learning_rate": 2.9999973245163716e-05,
       "loss": 0.7413,
       "step": 130
     },
     {
       "epoch": 6.086956521739131,
-      "grad_norm": 1.9068934917449951,
+      "grad_norm": 1.9071016311645508,
       "learning_rate": 2.99999524358576e-05,
-      "loss": 0.7653,
+      "loss": 0.7655,
       "step": 140
     },
     {
       "epoch": 6.521739130434782,
-      "grad_norm": 1.121341586112976,
+      "grad_norm": 1.1219961643218994,
       "learning_rate": 2.9999925681049593e-05,
       "loss": 0.7855,
       "step": 150
     },
     {
       "epoch": 6.521739130434782,
-      "eval_loss": 0.7960079908370972,
-      "eval_runtime": 0.4677,
-      "eval_samples_per_second": 21.382,
-      "eval_steps_per_second": 21.382,
+      "eval_loss": 0.7960985898971558,
+      "eval_runtime": 0.5137,
+      "eval_samples_per_second": 19.465,
+      "eval_steps_per_second": 19.465,
       "step": 150
     },
     {
       "epoch": 6.956521739130435,
-      "grad_norm": 1.5398368835449219,
+      "grad_norm": 1.5404783487319946,
       "learning_rate": 2.9999892980750297e-05,
-      "loss": 0.6587,
+      "loss": 0.6588,
       "step": 160
     },
     {
       "epoch": 7.391304347826087,
-      "grad_norm": 1.346014142036438,
+      "grad_norm": 1.3464210033416748,
       "learning_rate": 2.9999854334972675e-05,
       "loss": 0.739,
       "step": 170
     },
     {
       "epoch": 7.826086956521739,
-      "grad_norm": 1.7263554334640503,
+      "grad_norm": 1.7264870405197144,
       "learning_rate": 2.999980974373204e-05,
-      "loss": 0.729,
+      "loss": 0.7292,
       "step": 180
     },
     {
       "epoch": 8.26086956521739,
-      "grad_norm": 1.5398695468902588,
+      "grad_norm": 1.5393368005752563,
       "learning_rate": 2.9999759207046075e-05,
-      "loss": 0.6246,
+      "loss": 0.6244,
       "step": 190
     },
     {
       "epoch": 8.695652173913043,
-      "grad_norm": 1.7423065900802612,
+      "grad_norm": 1.7342770099639893,
       "learning_rate": 2.9999702724934804e-05,
-      "loss": 0.6762,
+      "loss": 0.6764,
       "step": 200
     },
     {
       "epoch": 9.130434782608695,
-      "grad_norm": 1.0418452024459839,
+      "grad_norm": 1.0416548252105713,
       "learning_rate": 2.999964029742062e-05,
-      "loss": 0.6522,
+      "loss": 0.6521,
       "step": 210
     },
     {
       "epoch": 9.565217391304348,
-      "grad_norm": 1.2121878862380981,
+      "grad_norm": 1.2193585634231567,
       "learning_rate": 2.9999571924528263e-05,
-      "loss": 0.559,
+      "loss": 0.5592,
       "step": 220
     },
     {
       "epoch": 10.0,
-      "grad_norm": 1.5282926559448242,
+      "grad_norm": 1.5286610126495361,
       "learning_rate": 2.9999497606284837e-05,
-      "loss": 0.7558,
+      "loss": 0.756,
       "step": 230
     },
     {
       "epoch": 10.434782608695652,
-      "grad_norm": 1.4222936630249023,
+      "grad_norm": 1.427451252937317,
       "learning_rate": 2.9999417342719796e-05,
-      "loss": 0.7119,
+      "loss": 0.7118,
       "step": 240
     },
     {
       "epoch": 10.869565217391305,
-      "grad_norm": 0.9751431941986084,
+      "grad_norm": 0.9794074296951294,
       "learning_rate": 2.9999331133864956e-05,
-      "loss": 0.59,
+      "loss": 0.5897,
       "step": 250
     },
     {
       "epoch": 11.304347826086957,
-      "grad_norm": 1.194357991218567,
+      "grad_norm": 1.194082260131836,
       "learning_rate": 2.9999238979754485e-05,
-      "loss": 0.6547,
+      "loss": 0.6548,
       "step": 260
     },
     {
       "epoch": 11.73913043478261,
-      "grad_norm": 1.049717903137207,
+      "grad_norm": 1.0489437580108643,
       "learning_rate": 2.999914088042492e-05,
       "loss": 0.6474,
       "step": 270
     },
     {
       "epoch": 12.173913043478262,
-      "grad_norm": 1.3123618364334106,
+      "grad_norm": 1.3105570077896118,
       "learning_rate": 2.9999036835915132e-05,
       "loss": 0.5936,
       "step": 280
     },
     {
       "epoch": 12.608695652173914,
-      "grad_norm": 1.0834040641784668,
+      "grad_norm": 1.0830979347229004,
       "learning_rate": 2.9998926846266365e-05,
       "loss": 0.6328,
       "step": 290
     },
     {
       "epoch": 13.043478260869565,
-      "grad_norm": 1.3875536918640137,
+      "grad_norm": 1.3943493366241455,
       "learning_rate": 2.9998810911522213e-05,
-      "loss": 0.5807,
+      "loss": 0.5806,
       "step": 300
     },
     {
       "epoch": 13.043478260869565,
-      "eval_loss": 0.7317642569541931,
-      "eval_runtime": 0.6288,
-      "eval_samples_per_second": 15.902,
-      "eval_steps_per_second": 15.902,
+      "eval_loss": 0.7312111854553223,
+      "eval_runtime": 0.6239,
+      "eval_samples_per_second": 16.028,
+      "eval_steps_per_second": 16.028,
       "step": 300
     },
     {
       "epoch": 13.478260869565217,
-      "grad_norm": 1.779242753982544,
+      "grad_norm": 1.7821625471115112,
       "learning_rate": 2.9998689031728636e-05,
-      "loss": 0.5146,
+      "loss": 0.5143,
       "step": 310
     },
     {
       "epoch": 13.91304347826087,
-      "grad_norm": 1.5307729244232178,
+      "grad_norm": 1.5245319604873657,
       "learning_rate": 2.9998561206933938e-05,
-      "loss": 0.6497,
+      "loss": 0.6494,
       "step": 320
     },
     {
       "epoch": 14.347826086956522,
-      "grad_norm": 1.479602575302124,
+      "grad_norm": 1.479852557182312,
       "learning_rate": 2.9998427437188786e-05,
-      "loss": 0.5746,
+      "loss": 0.5744,
       "step": 330
     },
     {
       "epoch": 14.782608695652174,
-      "grad_norm": 1.3181294202804565,
+      "grad_norm": 1.318290114402771,
       "learning_rate": 2.99982877225462e-05,
       "loss": 0.6014,
       "step": 340
     },
     {
       "epoch": 15.217391304347826,
-      "grad_norm": 0.9750956892967224,
+      "grad_norm": 0.9775345921516418,
       "learning_rate": 2.9998142063061564e-05,
       "loss": 0.4988,
       "step": 350
     },
     {
       "epoch": 15.652173913043478,
-      "grad_norm": 1.6386020183563232,
+      "grad_norm": 1.642491340637207,
       "learning_rate": 2.9997990458792603e-05,
-      "loss": 0.5627,
+      "loss": 0.5628,
       "step": 360
     },
     {
       "epoch": 16.08695652173913,
-      "grad_norm": 1.635014533996582,
+      "grad_norm": 1.6377389430999756,
       "learning_rate": 2.9997832909799417e-05,
-      "loss": 0.6674,
+      "loss": 0.6675,
       "step": 370
     },
     {
       "epoch": 16.52173913043478,
-      "grad_norm": 0.9534231424331665,
+      "grad_norm": 0.9537637829780579,
       "learning_rate": 2.9997669416144452e-05,
-      "loss": 0.513,
+      "loss": 0.5133,
       "step": 380
     },
     {
       "epoch": 16.956521739130434,
-      "grad_norm": 0.9356161952018738,
+      "grad_norm": 0.9352350234985352,
       "learning_rate": 2.999749997789251e-05,
-      "loss": 0.5795,
+      "loss": 0.5792,
       "step": 390
     },
     {
       "epoch": 17.391304347826086,
-      "grad_norm": 1.112004280090332,
+      "grad_norm": 1.112601399421692,
       "learning_rate": 2.9997324595110743e-05,
-      "loss": 0.518,
+      "loss": 0.5175,
       "step": 400
     },
     {
       "epoch": 17.82608695652174,
-      "grad_norm": 1.2841087579727173,
+      "grad_norm": 1.284747838973999,
       "learning_rate": 2.9997143267868683e-05,
-      "loss": 0.5878,
+      "loss": 0.5879,
       "step": 410
     },
     {
       "epoch": 18.26086956521739,
-      "grad_norm": 1.1624560356140137,
+      "grad_norm": 1.1622997522354126,
       "learning_rate": 2.9996955996238192e-05,
-      "loss": 0.5056,
+      "loss": 0.5061,
       "step": 420
     },
     {
       "epoch": 18.695652173913043,
-      "grad_norm": 1.2015905380249023,
+      "grad_norm": 1.1977813243865967,
       "learning_rate": 2.9996762780293503e-05,
-      "loss": 0.5317,
+      "loss": 0.5318,
       "step": 430
     },
     {
       "epoch": 19.130434782608695,
-      "grad_norm": 1.2141062021255493,
+      "grad_norm": 1.2117533683776855,
       "learning_rate": 2.9996563620111197e-05,
-      "loss": 0.5335,
+      "loss": 0.5336,
       "step": 440
     },
     {
       "epoch": 19.565217391304348,
-      "grad_norm": 1.4245619773864746,
+      "grad_norm": 1.4293084144592285,
       "learning_rate": 2.9996358515770218e-05,
-      "loss": 0.5677,
+      "loss": 0.5678,
       "step": 450
     },
     {
       "epoch": 19.565217391304348,
-      "eval_loss": 0.7168648838996887,
-      "eval_runtime": 0.6094,
-      "eval_samples_per_second": 16.41,
-      "eval_steps_per_second": 16.41,
+      "eval_loss": 0.7164609432220459,
+      "eval_runtime": 0.3868,
+      "eval_samples_per_second": 25.854,
+      "eval_steps_per_second": 25.854,
       "step": 450
     },
     {
       "epoch": 20.0,
-      "grad_norm": 2.1229870319366455,
+      "grad_norm": 2.1205103397369385,
       "learning_rate": 2.9996147467351856e-05,
-      "loss": 0.5146,
+      "loss": 0.5147,
       "step": 460
     },
     {
       "epoch": 20.434782608695652,
-      "grad_norm": 1.2757331132888794,
+      "grad_norm": 1.27629554271698,
       "learning_rate": 2.9995930474939773e-05,
-      "loss": 0.4783,
+      "loss": 0.4782,
       "step": 470
     },
     {
       "epoch": 20.869565217391305,
-      "grad_norm": 1.4738444089889526,
+      "grad_norm": 1.4705679416656494,
       "learning_rate": 2.9995707538619975e-05,
-      "loss": 0.5707,
+      "loss": 0.5705,
       "step": 480
     },
     {
       "epoch": 21.304347826086957,
-      "grad_norm": 1.3224209547042847,
+      "grad_norm": 1.3248465061187744,
       "learning_rate": 2.9995478658480822e-05,
-      "loss": 0.5163,
+      "loss": 0.5161,
       "step": 490
     },
     {
       "epoch": 21.73913043478261,
-      "grad_norm": 1.2427003383636475,
+      "grad_norm": 1.2424507141113281,
       "learning_rate": 2.9995243834613043e-05,
-      "loss": 0.5206,
+      "loss": 0.5204,
       "step": 500
     },
     {
       "epoch": 22.17391304347826,
-      "grad_norm": 1.7819838523864746,
+      "grad_norm": 1.783718228340149,
       "learning_rate": 2.9995003067109707e-05,
-      "loss": 0.4836,
+      "loss": 0.4837,
       "step": 510
     },
     {
       "epoch": 22.608695652173914,
-      "grad_norm": 1.531520128250122,
+      "grad_norm": 1.539941430091858,
       "learning_rate": 2.9994756356066246e-05,
-      "loss": 0.5615,
+      "loss": 0.5617,
       "step": 520
     },
     {
       "epoch": 23.043478260869566,
-      "grad_norm": 1.7395285367965698,
+      "grad_norm": 1.7417247295379639,
       "learning_rate": 2.999450370158046e-05,
       "loss": 0.4929,
       "step": 530
     },
     {
       "epoch": 23.47826086956522,
-      "grad_norm": 1.3100509643554688,
+      "grad_norm": 1.3086825609207153,
       "learning_rate": 2.9994245103752478e-05,
       "loss": 0.4383,
       "step": 540
     },
     {
       "epoch": 23.91304347826087,
-      "grad_norm": 1.2314648628234863,
+      "grad_norm": 1.2279788255691528,
       "learning_rate": 2.999398056268481e-05,
-      "loss": 0.5265,
+      "loss": 0.5266,
       "step": 550
     },
     {
       "epoch": 24.347826086956523,
-      "grad_norm": 1.4086766242980957,
+      "grad_norm": 1.40565025806427,
       "learning_rate": 2.9993710078482306e-05,
-      "loss": 0.5204,
+      "loss": 0.5207,
       "step": 560
     },
     {
       "epoch": 24.782608695652176,
-      "grad_norm": 0.9533238410949707,
+      "grad_norm": 0.9529893398284912,
       "learning_rate": 2.9993433651252185e-05,
-      "loss": 0.4432,
+      "loss": 0.4429,
       "step": 570
     },
     {
       "epoch": 25.217391304347824,
-      "grad_norm": 1.7112679481506348,
+      "grad_norm": 1.7137924432754517,
       "learning_rate": 2.9993151281104006e-05,
-      "loss": 0.5326,
+      "loss": 0.5325,
       "step": 580
     },
     {
       "epoch": 25.652173913043477,
-      "grad_norm": 1.1396631002426147,
+      "grad_norm": 1.1403567790985107,
       "learning_rate": 2.9992862968149695e-05,
-      "loss": 0.4736,
+      "loss": 0.4734,
       "step": 590
     },
     {
       "epoch": 26.08695652173913,
-      "grad_norm": 1.1695138216018677,
+      "grad_norm": 1.1702338457107544,
       "learning_rate": 2.9992568712503533e-05,
-      "loss": 0.4608,
+      "loss": 0.4609,
       "step": 600
     },
     {
       "epoch": 26.08695652173913,
-      "eval_loss": 0.7204726934432983,
-      "eval_runtime": 0.6002,
-      "eval_samples_per_second": 16.662,
-      "eval_steps_per_second": 16.662,
+      "eval_loss": 0.7209551334381104,
+      "eval_runtime": 0.4345,
+      "eval_samples_per_second": 23.015,
+      "eval_steps_per_second": 23.015,
       "step": 600
+    },
+    {
+      "epoch": 26.52173913043478,
+      "grad_norm": 1.076659917831421,
+      "learning_rate": 2.9992268514282142e-05,
+      "loss": 0.5118,
+      "step": 610
+    },
+    {
+      "epoch": 26.956521739130434,
+      "grad_norm": 1.320320963859558,
+      "learning_rate": 2.999196237360452e-05,
+      "loss": 0.4315,
+      "step": 620
+    },
+    {
+      "epoch": 27.391304347826086,
+      "grad_norm": 1.1899269819259644,
+      "learning_rate": 2.9991650290592016e-05,
+      "loss": 0.476,
+      "step": 630
+    },
+    {
+      "epoch": 27.82608695652174,
+      "grad_norm": 1.3601988554000854,
+      "learning_rate": 2.999133226536832e-05,
+      "loss": 0.5014,
+      "step": 640
+    },
+    {
+      "epoch": 28.26086956521739,
+      "grad_norm": 1.380045771598816,
+      "learning_rate": 2.9991008298059493e-05,
+      "loss": 0.4115,
+      "step": 650
+    },
+    {
+      "epoch": 28.695652173913043,
+      "grad_norm": 1.549436092376709,
+      "learning_rate": 2.9990678388793944e-05,
+      "loss": 0.5072,
+      "step": 660
+    },
+    {
+      "epoch": 29.130434782608695,
+      "grad_norm": 1.2787128686904907,
+      "learning_rate": 2.999034253770244e-05,
+      "loss": 0.4353,
+      "step": 670
+    },
+    {
+      "epoch": 29.565217391304348,
+      "grad_norm": 1.0582035779953003,
+      "learning_rate": 2.9990000744918097e-05,
+      "loss": 0.4715,
+      "step": 680
+    },
+    {
+      "epoch": 30.0,
+      "grad_norm": 2.4310319423675537,
+      "learning_rate": 2.9989653010576392e-05,
+      "loss": 0.4152,
+      "step": 690
+    },
+    {
+      "epoch": 30.434782608695652,
+      "grad_norm": 1.2908227443695068,
+      "learning_rate": 2.9989299334815158e-05,
+      "loss": 0.478,
+      "step": 700
+    },
+    {
+      "epoch": 30.869565217391305,
+      "grad_norm": 1.5819599628448486,
+      "learning_rate": 2.9988939717774578e-05,
+      "loss": 0.4132,
+      "step": 710
+    },
+    {
+      "epoch": 31.304347826086957,
+      "grad_norm": 0.8957367539405823,
+      "learning_rate": 2.9988574159597194e-05,
+      "loss": 0.4261,
+      "step": 720
+    },
+    {
+      "epoch": 31.73913043478261,
+      "grad_norm": 1.5984803438186646,
+      "learning_rate": 2.9988202660427907e-05,
+      "loss": 0.4842,
+      "step": 730
+    },
+    {
+      "epoch": 32.17391304347826,
+      "grad_norm": 1.1771948337554932,
+      "learning_rate": 2.9987825220413958e-05,
+      "loss": 0.4401,
+      "step": 740
+    },
+    {
+      "epoch": 32.608695652173914,
+      "grad_norm": 1.7258983850479126,
+      "learning_rate": 2.998744183970496e-05,
+      "loss": 0.4753,
+      "step": 750
+    },
+    {
+      "epoch": 32.608695652173914,
+      "eval_loss": 0.7315500974655151,
+      "eval_runtime": 0.4171,
+      "eval_samples_per_second": 23.973,
+      "eval_steps_per_second": 23.973,
+      "step": 750
     }
   ],
   "logging_steps": 10,
@@ -478,7 +591,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 1.5363065839583232e+16,
+  "total_flos": 1.9231766387195904e+16,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null

last-checkpoint/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a4e26a07b48ab9f4eeb67e7d951dfa1f71b1fc6e6c955372cec1a3c50fa8b9a2
+oid sha256:0c84fd1341bca21f546fb34f0984c2737ed54e8dbbe2ce01666781675b951db1
 size 5368