Training in progress, step 500, checkpoint

checkpoint-500/adapter_config.json

@@ -23,13 +23,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "o_proj",
-    "down_proj",
     "gate_proj",
     "k_proj",
+    "o_proj",
+    "up_proj",
     "q_proj",
-    "v_proj",
-    "up_proj"
+    "down_proj",
+    "v_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

checkpoint-500/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:af643e1258b1a0a26f4b1ae93e9cd05e3d41c90bf9e0da7b66d3fcb036bc5f59
+oid sha256:b88d46bc58ec7d7f7f41ea89e878c0553759853dfb81372d729a94f29745f7a0
 size 145287696

checkpoint-500/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a16a0a3418cc3bfc9c02d50a9fc9737cbaa9b35d8bd5ddd509621e46d71984b7
+oid sha256:48258933f70a2d23d72c37994f6598bd88dd0686bdc68bacd273e934a256224f
 size 290833618

checkpoint-500/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ae8708c0d3e4790c415a8daad07efc7d696d468ca1f90f0b4e30cc974ceff368
+oid sha256:b1dc1f142c3170be23a0047b17b034f9477746f310baa812792a3bf136624c89
 size 14244

checkpoint-500/trainer_state.json

@@ -10,432 +10,432 @@
   "log_history": [
     {
       "epoch": 0.010869565217391304,
-      "grad_norm": 2.120478868484497,
+      "grad_norm": 2.4570870399475098,
       "learning_rate": 4.981884057971015e-05,
-      "loss": 2.4608,
+      "loss": 2.3612,
       "step": 10
     },
     {
       "epoch": 0.021739130434782608,
-      "grad_norm": 1.4271776676177979,
+      "grad_norm": 1.2042421102523804,
       "learning_rate": 4.963768115942029e-05,
-      "loss": 1.2741,
+      "loss": 1.1915,
       "step": 20
     },
     {
       "epoch": 0.03260869565217391,
-      "grad_norm": 0.9902746677398682,
+      "grad_norm": 1.2019174098968506,
       "learning_rate": 4.945652173913044e-05,
-      "loss": 0.9403,
+      "loss": 0.8992,
       "step": 30
     },
     {
       "epoch": 0.043478260869565216,
-      "grad_norm": 1.0730223655700684,
+      "grad_norm": 1.0546120405197144,
       "learning_rate": 4.9275362318840584e-05,
-      "loss": 0.7517,
+      "loss": 0.7276,
       "step": 40
     },
     {
       "epoch": 0.05434782608695652,
-      "grad_norm": 0.8592391610145569,
+      "grad_norm": 0.754091203212738,
       "learning_rate": 4.909420289855073e-05,
-      "loss": 0.6695,
+      "loss": 0.6419,
       "step": 50
     },
     {
       "epoch": 0.05434782608695652,
-      "eval_loss": 0.8473024964332581,
-      "eval_runtime": 10.3597,
-      "eval_samples_per_second": 46.334,
-      "eval_steps_per_second": 2.896,
+      "eval_loss": 0.7905128598213196,
+      "eval_runtime": 10.8418,
+      "eval_samples_per_second": 44.273,
+      "eval_steps_per_second": 2.767,
       "step": 50
     },
     {
       "epoch": 0.06521739130434782,
-      "grad_norm": 0.7817952632904053,
+      "grad_norm": 0.7751966714859009,
       "learning_rate": 4.891304347826087e-05,
-      "loss": 0.6325,
+      "loss": 0.5971,
       "step": 60
     },
     {
       "epoch": 0.07608695652173914,
-      "grad_norm": 0.8163846135139465,
+      "grad_norm": 0.6874057650566101,
       "learning_rate": 4.873188405797102e-05,
-      "loss": 0.5994,
+      "loss": 0.608,
       "step": 70
     },
     {
       "epoch": 0.08695652173913043,
-      "grad_norm": 0.7358843088150024,
+      "grad_norm": 0.7145748734474182,
       "learning_rate": 4.855072463768116e-05,
-      "loss": 0.6158,
+      "loss": 0.6111,
       "step": 80
     },
     {
       "epoch": 0.09782608695652174,
-      "grad_norm": 0.7856508493423462,
+      "grad_norm": 0.8841484189033508,
       "learning_rate": 4.836956521739131e-05,
-      "loss": 0.5983,
+      "loss": 0.5848,
       "step": 90
     },
     {
       "epoch": 0.10869565217391304,
-      "grad_norm": 0.8615956902503967,
+      "grad_norm": 0.7937784790992737,
       "learning_rate": 4.818840579710145e-05,
-      "loss": 0.5858,
+      "loss": 0.5699,
       "step": 100
     },
     {
       "epoch": 0.10869565217391304,
-      "eval_loss": 0.7398766875267029,
-      "eval_runtime": 10.4317,
-      "eval_samples_per_second": 46.014,
-      "eval_steps_per_second": 2.876,
+      "eval_loss": 0.6956210732460022,
+      "eval_runtime": 10.863,
+      "eval_samples_per_second": 44.187,
+      "eval_steps_per_second": 2.762,
       "step": 100
     },
     {
       "epoch": 0.11956521739130435,
-      "grad_norm": 0.7915664315223694,
+      "grad_norm": 0.7818441987037659,
       "learning_rate": 4.80072463768116e-05,
-      "loss": 0.6228,
+      "loss": 0.5946,
       "step": 110
     },
     {
       "epoch": 0.13043478260869565,
-      "grad_norm": 0.8432077765464783,
+      "grad_norm": 0.8666340112686157,
       "learning_rate": 4.782608695652174e-05,
-      "loss": 0.5649,
+      "loss": 0.5469,
       "step": 120
     },
     {
       "epoch": 0.14130434782608695,
-      "grad_norm": 0.8230385780334473,
+      "grad_norm": 0.7637468576431274,
       "learning_rate": 4.764492753623189e-05,
-      "loss": 0.5368,
+      "loss": 0.5307,
       "step": 130
     },
     {
       "epoch": 0.15217391304347827,
-      "grad_norm": 0.8497728109359741,
+      "grad_norm": 0.8282362222671509,
       "learning_rate": 4.746376811594203e-05,
-      "loss": 0.5447,
+      "loss": 0.5312,
       "step": 140
     },
     {
       "epoch": 0.16304347826086957,
-      "grad_norm": 0.9394277930259705,
+      "grad_norm": 0.9675197601318359,
       "learning_rate": 4.7282608695652177e-05,
-      "loss": 0.5515,
+      "loss": 0.5486,
       "step": 150
     },
     {
       "epoch": 0.16304347826086957,
-      "eval_loss": 0.694922149181366,
-      "eval_runtime": 10.4348,
-      "eval_samples_per_second": 46.0,
-      "eval_steps_per_second": 2.875,
+      "eval_loss": 0.6567058563232422,
+      "eval_runtime": 10.8376,
+      "eval_samples_per_second": 44.29,
+      "eval_steps_per_second": 2.768,
       "step": 150
     },
     {
       "epoch": 0.17391304347826086,
-      "grad_norm": 0.7665875554084778,
+      "grad_norm": 0.7035924792289734,
       "learning_rate": 4.710144927536232e-05,
-      "loss": 0.5632,
+      "loss": 0.5553,
       "step": 160
     },
     {
       "epoch": 0.18478260869565216,
-      "grad_norm": 0.8912480473518372,
+      "grad_norm": 0.8957257866859436,
       "learning_rate": 4.6920289855072464e-05,
-      "loss": 0.5259,
+      "loss": 0.54,
       "step": 170
     },
     {
       "epoch": 0.1956521739130435,
-      "grad_norm": 0.8959771990776062,
+      "grad_norm": 0.8544663190841675,
       "learning_rate": 4.673913043478261e-05,
-      "loss": 0.5307,
+      "loss": 0.55,
       "step": 180
     },
     {
       "epoch": 0.20652173913043478,
-      "grad_norm": 0.8595960140228271,
+      "grad_norm": 0.7599456310272217,
       "learning_rate": 4.655797101449276e-05,
-      "loss": 0.5167,
+      "loss": 0.5102,
       "step": 190
     },
     {
       "epoch": 0.21739130434782608,
-      "grad_norm": 0.8418941497802734,
+      "grad_norm": 0.9151259064674377,
       "learning_rate": 4.63768115942029e-05,
-      "loss": 0.5106,
+      "loss": 0.5372,
       "step": 200
     },
     {
       "epoch": 0.21739130434782608,
-      "eval_loss": 0.6705857515335083,
-      "eval_runtime": 10.4259,
-      "eval_samples_per_second": 46.039,
-      "eval_steps_per_second": 2.877,
+      "eval_loss": 0.6319016218185425,
+      "eval_runtime": 10.8237,
+      "eval_samples_per_second": 44.347,
+      "eval_steps_per_second": 2.772,
       "step": 200
     },
     {
       "epoch": 0.22826086956521738,
-      "grad_norm": 0.8937019109725952,
+      "grad_norm": 0.911018431186676,
       "learning_rate": 4.6195652173913046e-05,
-      "loss": 0.5329,
+      "loss": 0.5354,
       "step": 210
     },
     {
       "epoch": 0.2391304347826087,
-      "grad_norm": 0.9121237397193909,
+      "grad_norm": 0.7792191505432129,
       "learning_rate": 4.601449275362319e-05,
-      "loss": 0.5341,
+      "loss": 0.5297,
       "step": 220
     },
     {
       "epoch": 0.25,
-      "grad_norm": 0.8253990411758423,
+      "grad_norm": 0.7631202340126038,
       "learning_rate": 4.5833333333333334e-05,
-      "loss": 0.5328,
+      "loss": 0.5258,
       "step": 230
     },
     {
       "epoch": 0.2608695652173913,
-      "grad_norm": 0.9218540787696838,
+      "grad_norm": 0.7436497211456299,
       "learning_rate": 4.565217391304348e-05,
-      "loss": 0.4706,
+      "loss": 0.4849,
       "step": 240
     },
     {
       "epoch": 0.2717391304347826,
-      "grad_norm": 0.9071463346481323,
+      "grad_norm": 0.8372677564620972,
       "learning_rate": 4.547101449275363e-05,
-      "loss": 0.5012,
+      "loss": 0.5131,
       "step": 250
     },
     {
       "epoch": 0.2717391304347826,
-      "eval_loss": 0.6487293839454651,
-      "eval_runtime": 10.4286,
-      "eval_samples_per_second": 46.027,
-      "eval_steps_per_second": 2.877,
+      "eval_loss": 0.6155125498771667,
+      "eval_runtime": 10.8327,
+      "eval_samples_per_second": 44.31,
+      "eval_steps_per_second": 2.769,
       "step": 250
     },
     {
       "epoch": 0.2826086956521739,
-      "grad_norm": 0.9131689071655273,
+      "grad_norm": 0.8056390285491943,
       "learning_rate": 4.528985507246377e-05,
-      "loss": 0.4819,
+      "loss": 0.4675,
       "step": 260
     },
     {
       "epoch": 0.29347826086956524,
-      "grad_norm": 0.8460387587547302,
+      "grad_norm": 0.8015912175178528,
       "learning_rate": 4.5108695652173916e-05,
-      "loss": 0.4951,
+      "loss": 0.4942,
       "step": 270
     },
     {
       "epoch": 0.30434782608695654,
-      "grad_norm": 0.8639910817146301,
+      "grad_norm": 0.8316887021064758,
       "learning_rate": 4.492753623188406e-05,
-      "loss": 0.4941,
+      "loss": 0.5025,
       "step": 280
     },
     {
       "epoch": 0.31521739130434784,
-      "grad_norm": 1.0238360166549683,
+      "grad_norm": 0.8440260887145996,
       "learning_rate": 4.4746376811594203e-05,
-      "loss": 0.4844,
+      "loss": 0.5112,
       "step": 290
     },
     {
       "epoch": 0.32608695652173914,
-      "grad_norm": 0.8217597603797913,
+      "grad_norm": 0.8523954153060913,
       "learning_rate": 4.456521739130435e-05,
-      "loss": 0.497,
+      "loss": 0.4973,
       "step": 300
     },
     {
       "epoch": 0.32608695652173914,
-      "eval_loss": 0.6351093053817749,
-      "eval_runtime": 10.4278,
-      "eval_samples_per_second": 46.031,
-      "eval_steps_per_second": 2.877,
+      "eval_loss": 0.6024141907691956,
+      "eval_runtime": 10.839,
+      "eval_samples_per_second": 44.284,
+      "eval_steps_per_second": 2.768,
       "step": 300
     },
     {
       "epoch": 0.33695652173913043,
-      "grad_norm": 0.8742851614952087,
+      "grad_norm": 0.8427216410636902,
       "learning_rate": 4.438405797101449e-05,
-      "loss": 0.5134,
+      "loss": 0.5271,
       "step": 310
     },
     {
       "epoch": 0.34782608695652173,
-      "grad_norm": 0.8606206774711609,
+      "grad_norm": 0.8126628398895264,
       "learning_rate": 4.4202898550724645e-05,
-      "loss": 0.4577,
+      "loss": 0.4608,
       "step": 320
     },
     {
       "epoch": 0.358695652173913,
-      "grad_norm": 1.030929446220398,
+      "grad_norm": 0.9947460293769836,
       "learning_rate": 4.4021739130434786e-05,
-      "loss": 0.4624,
+      "loss": 0.4762,
       "step": 330
     },
     {
       "epoch": 0.3695652173913043,
-      "grad_norm": 0.8211489915847778,
+      "grad_norm": 0.7744721174240112,
       "learning_rate": 4.384057971014493e-05,
-      "loss": 0.4942,
+      "loss": 0.4993,
       "step": 340
     },
     {
       "epoch": 0.3804347826086957,
-      "grad_norm": 1.0272365808486938,
+      "grad_norm": 0.9483833909034729,
       "learning_rate": 4.365942028985507e-05,
-      "loss": 0.4923,
+      "loss": 0.497,
       "step": 350
     },
     {
       "epoch": 0.3804347826086957,
-      "eval_loss": 0.6294742226600647,
-      "eval_runtime": 10.439,
-      "eval_samples_per_second": 45.981,
-      "eval_steps_per_second": 2.874,
+      "eval_loss": 0.596432626247406,
+      "eval_runtime": 10.8191,
+      "eval_samples_per_second": 44.366,
+      "eval_steps_per_second": 2.773,
       "step": 350
     },
     {
       "epoch": 0.391304347826087,
-      "grad_norm": 0.9275714159011841,
+      "grad_norm": 0.7573915123939514,
       "learning_rate": 4.347826086956522e-05,
-      "loss": 0.4784,
+      "loss": 0.469,
       "step": 360
     },
     {
       "epoch": 0.40217391304347827,
-      "grad_norm": 0.9278607368469238,
+      "grad_norm": 0.8013887405395508,
       "learning_rate": 4.329710144927536e-05,
-      "loss": 0.465,
+      "loss": 0.4704,
       "step": 370
     },
     {
       "epoch": 0.41304347826086957,
-      "grad_norm": 1.009921669960022,
+      "grad_norm": 0.8412826657295227,
       "learning_rate": 4.3115942028985515e-05,
-      "loss": 0.4776,
+      "loss": 0.5033,
       "step": 380
     },
     {
       "epoch": 0.42391304347826086,
-      "grad_norm": 0.9694483280181885,
+      "grad_norm": 0.8203413486480713,
       "learning_rate": 4.2934782608695655e-05,
-      "loss": 0.4956,
+      "loss": 0.5063,
       "step": 390
     },
     {
       "epoch": 0.43478260869565216,
-      "grad_norm": 0.8650519251823425,
+      "grad_norm": 0.7952508926391602,
       "learning_rate": 4.27536231884058e-05,
-      "loss": 0.4607,
+      "loss": 0.4706,
       "step": 400
     },
     {
       "epoch": 0.43478260869565216,
-      "eval_loss": 0.6202508807182312,
-      "eval_runtime": 10.4217,
-      "eval_samples_per_second": 46.058,
-      "eval_steps_per_second": 2.879,
+      "eval_loss": 0.5875406861305237,
+      "eval_runtime": 10.8328,
+      "eval_samples_per_second": 44.31,
+      "eval_steps_per_second": 2.769,
       "step": 400
     },
     {
       "epoch": 0.44565217391304346,
-      "grad_norm": 1.0009912252426147,
+      "grad_norm": 0.9020015597343445,
       "learning_rate": 4.257246376811594e-05,
-      "loss": 0.4695,
+      "loss": 0.4786,
       "step": 410
     },
     {
       "epoch": 0.45652173913043476,
-      "grad_norm": 0.9651597738265991,
+      "grad_norm": 0.8588363528251648,
       "learning_rate": 4.239130434782609e-05,
-      "loss": 0.4733,
+      "loss": 0.482,
       "step": 420
     },
     {
       "epoch": 0.4673913043478261,
-      "grad_norm": 0.9021621346473694,
+      "grad_norm": 0.816198468208313,
       "learning_rate": 4.221014492753623e-05,
-      "loss": 0.4543,
+      "loss": 0.4683,
       "step": 430
     },
     {
       "epoch": 0.4782608695652174,
-      "grad_norm": 0.9730608463287354,
+      "grad_norm": 0.8713563680648804,
       "learning_rate": 4.202898550724638e-05,
-      "loss": 0.4461,
+      "loss": 0.4422,
       "step": 440
     },
     {
       "epoch": 0.4891304347826087,
-      "grad_norm": 0.9395557641983032,
+      "grad_norm": 0.8633397817611694,
       "learning_rate": 4.1847826086956525e-05,
-      "loss": 0.4448,
+      "loss": 0.45,
       "step": 450
     },
     {
       "epoch": 0.4891304347826087,
-      "eval_loss": 0.6122747659683228,
-      "eval_runtime": 10.4358,
-      "eval_samples_per_second": 45.995,
-      "eval_steps_per_second": 2.875,
+      "eval_loss": 0.5812788605690002,
+      "eval_runtime": 10.8048,
+      "eval_samples_per_second": 44.425,
+      "eval_steps_per_second": 2.777,
       "step": 450
     },
     {
       "epoch": 0.5,
-      "grad_norm": 0.9062746167182922,
+      "grad_norm": 0.8244722485542297,
       "learning_rate": 4.166666666666667e-05,
-      "loss": 0.4653,
+      "loss": 0.4911,
       "step": 460
     },
     {
       "epoch": 0.5108695652173914,
-      "grad_norm": 0.9474561214447021,
+      "grad_norm": 0.8201929926872253,
       "learning_rate": 4.148550724637681e-05,
-      "loss": 0.4571,
+      "loss": 0.4611,
       "step": 470
     },
     {
       "epoch": 0.5217391304347826,
-      "grad_norm": 0.9197052717208862,
+      "grad_norm": 0.9646862149238586,
       "learning_rate": 4.130434782608696e-05,
-      "loss": 0.4416,
+      "loss": 0.4738,
       "step": 480
     },
     {
       "epoch": 0.532608695652174,
-      "grad_norm": 0.9986901879310608,
+      "grad_norm": 0.9149807095527649,
       "learning_rate": 4.11231884057971e-05,
-      "loss": 0.4446,
+      "loss": 0.4603,
       "step": 490
     },
     {
       "epoch": 0.5434782608695652,
-      "grad_norm": 0.9315592050552368,
+      "grad_norm": 0.8054277896881104,
       "learning_rate": 4.094202898550725e-05,
-      "loss": 0.4501,
+      "loss": 0.466,
       "step": 500
     },
     {
       "epoch": 0.5434782608695652,
-      "eval_loss": 0.6061260104179382,
-      "eval_runtime": 10.4306,
-      "eval_samples_per_second": 46.018,
-      "eval_steps_per_second": 2.876,
+      "eval_loss": 0.573806881904602,
+      "eval_runtime": 10.7906,
+      "eval_samples_per_second": 44.483,
+      "eval_steps_per_second": 2.78,
       "step": 500
     }
   ],
@@ -456,7 +456,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 4.418132187414528e+16,
+  "total_flos": 4.386145963514266e+16,
   "train_batch_size": 16,
   "trial_name": null,
   "trial_params": null

checkpoint-500/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a3712c5fdd6e5241e8ea26ae62b28612fc514ec88e8a03bc6c1b3a46acfc611a
+oid sha256:81786d634f359bee447d9c2dad023b92502900c87e9208470ed326b1204c5b1a
 size 5816