Training in progress, epoch 9, checkpoint

checkpoint-32391/README.md

@@ -199,4 +199,4 @@ Carbon emissions can be estimated using the [Machine Learning Impact calculator]
 [More Information Needed]
 ### Framework versions
 
-- PEFT 0.15.2
+- PEFT 0.14.0

checkpoint-32391/adapter_config.json

@@ -3,7 +3,6 @@
   "auto_mapping": null,
   "base_model_name_or_path": "deepseek-ai/deepseek-coder-1.3b-base",
   "bias": "none",
-  "corda_config": null,
   "eva_config": null,
   "exclude_modules": null,
   "fan_in_fan_out": false,
@@ -24,11 +23,10 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "q_proj",
-    "v_proj"
+    "v_proj",
+    "q_proj"
   ],
   "task_type": "CAUSAL_LM",
-  "trainable_token_indices": null,
   "use_dora": false,
   "use_rslora": false
 }

checkpoint-32391/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:847ff5b7095ec28f102fca048038d2b3aa93c56dc5e9de622c5f10b11a59ac0f
+oid sha256:fa77979c1338f0e5aaadde6fcc4c8d3546a9776c0300b343cdfc819a7203b165
 size 268636736

checkpoint-32391/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1e2abee755745546134395a29f5f118d29ac7b2477e3646abdd3907b09c6312f
-size 12664203
+oid sha256:a858bcc6a44ea96ab1945ddf226f3ce2766f800a01f06327d0a2630ceea2b09b
+size 12663802

checkpoint-32391/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:04f762123ef5f1c6bef4402cb0c4755573312d3042f074d785a0cef3b735f025
-size 16389
+oid sha256:76b32d09a1883ad1c756dcc344184725264d3a6b881c334469a2deb7bc865e1b
+size 15984

checkpoint-32391/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f9beb007dfd20d7de2b4adc1082e0aee6546737d3371ec89a7e979c117d409ae
-size 16389
+oid sha256:d9869aa141ea6e4d7356b2e53065da5698b2016a95015b3b066b99931f0cc04f
+size 15984

checkpoint-32391/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bf7419293dca88b57e77134603d1871918dbd66e478f54e3f7cab06f7bfe775d
-size 16389
+oid sha256:046406ffe03a772d5f2323d116cc759fd0519bed354740f6eabfff368049c076
+size 15984

checkpoint-32391/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:41c00894de21342973a37a344f76bff3f037a72f15dac50db610054be44e56c0
-size 16389
+oid sha256:3439b8f870442ae79e732f10dfcc3fa7379ff6477fb60bc395f1677d9e9bc6d7
+size 15984

checkpoint-32391/rng_state_4.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9c3d2b76a6df5659b4bca611f1cd7d5233e44a3977daace77af4e600560f5c98
-size 16389
+oid sha256:c0a872c73196b795a3d2a172f8c0b57f477b546f74fc24b1384275c69bf42da7
+size 15984

checkpoint-32391/rng_state_5.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:59e8516e4e604c97a2ffbcebfbe57dd2c042c3a8e879c0bfa03599adddbdd1af
-size 16389
+oid sha256:7bbe7a052a0e5a7234bf6bce6821051a71cf85e4b561b98b5fbe2110f0434bdf
+size 15984

checkpoint-32391/rng_state_6.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7a0a12c1b31e99194d757aca9cfd909b2e2d05074fdedd24724ee09772b8aa00
-size 16389
+oid sha256:ec92094cfa8a5cd3e8602ab3b105cbb06d93f04acd00270ba605daf32c02beaf
+size 15984

checkpoint-32391/rng_state_7.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3078852551d2813d4f0ae61da508ab9aa38f63962a15b2d60052cb69724f4007
-size 16389
+oid sha256:499ec2a203ee5c5c50787874545bdc219bffabfe56927e72cb6ea1c6f20b998e
+size 15984

checkpoint-32391/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6659fae724bd5a86aebda0bd36b55dca38a567007abee4b02611a3fd11eff0ed
-size 1465
+oid sha256:970c1967a44b52c9683d7a838c466a40332395dcc79a19cecd04df4b17d1e9a8
+size 1064

checkpoint-32391/trainer_state.json

@@ -1,5 +1,4 @@
 {
-  "best_global_step": null,
   "best_metric": null,
   "best_model_checkpoint": null,
   "epoch": 9.0,
@@ -11,802 +10,802 @@
   "log_history": [
     {
       "epoch": 0.13892747985551543,
-      "grad_norm": 0.6735790967941284,
-      "learning_rate": 0.0007907937390015746,
-      "loss": 0.3742,
+      "grad_norm": 0.9573676586151123,
+      "learning_rate": 0.0007907566916736131,
+      "loss": 0.377,
       "step": 500
     },
     {
       "epoch": 0.2000555709919422,
-      "eval_loss": 0.2844131588935852,
-      "eval_runtime": 5.091,
-      "eval_samples_per_second": 98.212,
-      "eval_steps_per_second": 6.286,
+      "eval_loss": 0.28939542174339294,
+      "eval_runtime": 4.6518,
+      "eval_samples_per_second": 107.484,
+      "eval_steps_per_second": 6.879,
       "step": 720
     },
     {
       "epoch": 0.27785495971103086,
-      "grad_norm": 0.6214504241943359,
-      "learning_rate": 0.0007815319070112068,
-      "loss": 0.2859,
+      "grad_norm": 0.5571001768112183,
+      "learning_rate": 0.0007815133833472261,
+      "loss": 0.2824,
       "step": 1000
     },
     {
       "epoch": 0.4001111419838844,
-      "eval_loss": 0.2708466649055481,
-      "eval_runtime": 4.8349,
-      "eval_samples_per_second": 103.416,
-      "eval_steps_per_second": 6.619,
+      "eval_loss": 0.2604603171348572,
+      "eval_runtime": 4.3083,
+      "eval_samples_per_second": 116.054,
+      "eval_steps_per_second": 7.427,
       "step": 1440
     },
     {
       "epoch": 0.41678243956654626,
-      "grad_norm": 0.5057302117347717,
-      "learning_rate": 0.0007722700750208392,
-      "loss": 0.2679,
+      "grad_norm": 0.6729081273078918,
+      "learning_rate": 0.0007722515513568584,
+      "loss": 0.2627,
       "step": 1500
     },
     {
       "epoch": 0.5557099194220617,
-      "grad_norm": 0.623817503452301,
-      "learning_rate": 0.0007630082430304714,
-      "loss": 0.2527,
+      "grad_norm": 0.5200651288032532,
+      "learning_rate": 0.0007629897193664907,
+      "loss": 0.2488,
       "step": 2000
     },
     {
       "epoch": 0.6001667129758266,
-      "eval_loss": 0.24877774715423584,
-      "eval_runtime": 5.9656,
-      "eval_samples_per_second": 83.813,
-      "eval_steps_per_second": 5.364,
+      "eval_loss": 0.24864767491817474,
+      "eval_runtime": 3.8713,
+      "eval_samples_per_second": 129.157,
+      "eval_steps_per_second": 8.266,
       "step": 2160
     },
     {
       "epoch": 0.6946373992775771,
-      "grad_norm": 1.3911620378494263,
-      "learning_rate": 0.0007537649347040845,
-      "loss": 0.2441,
+      "grad_norm": 0.6645380258560181,
+      "learning_rate": 0.0007537464110401038,
+      "loss": 0.2424,
       "step": 2500
     },
     {
       "epoch": 0.8002222839677688,
-      "eval_loss": 0.2391277253627777,
-      "eval_runtime": 5.9872,
-      "eval_samples_per_second": 83.512,
-      "eval_steps_per_second": 5.345,
+      "eval_loss": 0.2434273064136505,
+      "eval_runtime": 4.2012,
+      "eval_samples_per_second": 119.013,
+      "eval_steps_per_second": 7.617,
       "step": 2880
     },
     {
       "epoch": 0.8335648791330925,
-      "grad_norm": 0.7922923564910889,
-      "learning_rate": 0.0007445216263776975,
-      "loss": 0.235,
+      "grad_norm": 0.8886778354644775,
+      "learning_rate": 0.000744484579049736,
+      "loss": 0.2332,
       "step": 3000
     },
     {
       "epoch": 0.972492358988608,
-      "grad_norm": 0.8539218902587891,
-      "learning_rate": 0.0007352597943873298,
-      "loss": 0.2331,
+      "grad_norm": 0.5990277528762817,
+      "learning_rate": 0.0007352227470593684,
+      "loss": 0.2307,
       "step": 3500
     },
     {
       "epoch": 1.000277854959711,
-      "eval_loss": 0.23887279629707336,
-      "eval_runtime": 5.6004,
-      "eval_samples_per_second": 89.279,
-      "eval_steps_per_second": 5.714,
+      "eval_loss": 0.23320984840393066,
+      "eval_runtime": 4.3777,
+      "eval_samples_per_second": 114.216,
+      "eval_steps_per_second": 7.31,
       "step": 3600
     },
     {
       "epoch": 1.1114198388441234,
-      "grad_norm": 0.6812190413475037,
-      "learning_rate": 0.0007259979623969622,
-      "loss": 0.2201,
+      "grad_norm": 0.9014330506324768,
+      "learning_rate": 0.0007259609150690007,
+      "loss": 0.2171,
       "step": 4000
     },
     {
       "epoch": 1.2003334259516532,
-      "eval_loss": 0.2331564724445343,
-      "eval_runtime": 5.2457,
-      "eval_samples_per_second": 95.317,
-      "eval_steps_per_second": 6.1,
+      "eval_loss": 0.23165521025657654,
+      "eval_runtime": 4.5533,
+      "eval_samples_per_second": 109.81,
+      "eval_steps_per_second": 7.028,
       "step": 4320
     },
     {
       "epoch": 1.2503473186996388,
-      "grad_norm": 0.6926279067993164,
-      "learning_rate": 0.0007167361304065945,
-      "loss": 0.2199,
+      "grad_norm": 0.9070354104042053,
+      "learning_rate": 0.000716699083078633,
+      "loss": 0.2174,
       "step": 4500
     },
     {
       "epoch": 1.3892747985551543,
-      "grad_norm": 0.857128918170929,
-      "learning_rate": 0.0007074742984162268,
-      "loss": 0.2157,
+      "grad_norm": 0.9023842215538025,
+      "learning_rate": 0.000707455774752246,
+      "loss": 0.215,
       "step": 5000
     },
     {
       "epoch": 1.4003889969435954,
-      "eval_loss": 0.22209732234477997,
-      "eval_runtime": 4.9003,
-      "eval_samples_per_second": 102.035,
-      "eval_steps_per_second": 6.53,
+      "eval_loss": 0.21996016800403595,
+      "eval_runtime": 4.2649,
+      "eval_samples_per_second": 117.235,
+      "eval_steps_per_second": 7.503,
       "step": 5040
     },
     {
       "epoch": 1.5282022784106695,
-      "grad_norm": 0.8350584506988525,
-      "learning_rate": 0.000698212466425859,
-      "loss": 0.2122,
+      "grad_norm": 0.8724926710128784,
+      "learning_rate": 0.0006981939427618783,
+      "loss": 0.2101,
       "step": 5500
     },
     {
       "epoch": 1.6004445679355377,
-      "eval_loss": 0.22165100276470184,
-      "eval_runtime": 4.8879,
-      "eval_samples_per_second": 102.294,
-      "eval_steps_per_second": 6.547,
+      "eval_loss": 0.21771642565727234,
+      "eval_runtime": 4.6518,
+      "eval_samples_per_second": 107.486,
+      "eval_steps_per_second": 6.879,
       "step": 5760
     },
     {
       "epoch": 1.667129758266185,
-      "grad_norm": 0.6862344145774841,
-      "learning_rate": 0.0006889691580994721,
-      "loss": 0.2158,
+      "grad_norm": 1.1229606866836548,
+      "learning_rate": 0.0006889321107715106,
+      "loss": 0.209,
       "step": 6000
     },
     {
       "epoch": 1.8005001389274797,
-      "eval_loss": 0.21265414357185364,
-      "eval_runtime": 5.1036,
-      "eval_samples_per_second": 97.97,
-      "eval_steps_per_second": 6.27,
+      "eval_loss": 0.2162657231092453,
+      "eval_runtime": 3.9728,
+      "eval_samples_per_second": 125.857,
+      "eval_steps_per_second": 8.055,
       "step": 6480
     },
     {
       "epoch": 1.8060572381217006,
-      "grad_norm": 0.64631587266922,
-      "learning_rate": 0.0006797073261091044,
-      "loss": 0.2069,
+      "grad_norm": 0.562541127204895,
+      "learning_rate": 0.000679670278781143,
+      "loss": 0.2053,
       "step": 6500
     },
     {
       "epoch": 1.9449847179772157,
-      "grad_norm": 0.6139649748802185,
-      "learning_rate": 0.0006704454941187368,
-      "loss": 0.2063,
+      "grad_norm": 0.9058884382247925,
+      "learning_rate": 0.0006704084467907753,
+      "loss": 0.2047,
       "step": 7000
     },
     {
       "epoch": 2.000555709919422,
-      "eval_loss": 0.2054276168346405,
-      "eval_runtime": 5.6047,
-      "eval_samples_per_second": 89.211,
-      "eval_steps_per_second": 5.71,
+      "eval_loss": 0.21751119196414948,
+      "eval_runtime": 3.8866,
+      "eval_samples_per_second": 128.646,
+      "eval_steps_per_second": 8.233,
       "step": 7200
     },
     {
       "epoch": 2.0839121978327313,
-      "grad_norm": 0.5134121179580688,
-      "learning_rate": 0.000661183662128369,
-      "loss": 0.1961,
+      "grad_norm": 0.5113864541053772,
+      "learning_rate": 0.0006611466148004076,
+      "loss": 0.1972,
       "step": 7500
     },
     {
       "epoch": 2.2006112809113643,
-      "eval_loss": 0.20968343317508698,
-      "eval_runtime": 4.9427,
-      "eval_samples_per_second": 101.16,
-      "eval_steps_per_second": 6.474,
+      "eval_loss": 0.21081140637397766,
+      "eval_runtime": 3.9021,
+      "eval_samples_per_second": 128.136,
+      "eval_steps_per_second": 8.201,
       "step": 7920
     },
     {
       "epoch": 2.222839677688247,
-      "grad_norm": 0.667497456073761,
-      "learning_rate": 0.0006519218301380013,
-      "loss": 0.198,
+      "grad_norm": 0.6523454785346985,
+      "learning_rate": 0.0006518847828100398,
+      "loss": 0.1968,
       "step": 8000
     },
     {
       "epoch": 2.361767157543762,
-      "grad_norm": 0.7320435643196106,
-      "learning_rate": 0.0006426599981476336,
-      "loss": 0.1942,
+      "grad_norm": 0.48720309138298035,
+      "learning_rate": 0.0006426414744836529,
+      "loss": 0.1934,
       "step": 8500
     },
     {
       "epoch": 2.4006668519033063,
-      "eval_loss": 0.201798677444458,
-      "eval_runtime": 5.0966,
-      "eval_samples_per_second": 98.105,
-      "eval_steps_per_second": 6.279,
+      "eval_loss": 0.20452378690242767,
+      "eval_runtime": 4.661,
+      "eval_samples_per_second": 107.272,
+      "eval_steps_per_second": 6.865,
       "step": 8640
     },
     {
       "epoch": 2.5006946373992776,
-      "grad_norm": 0.5457369685173035,
-      "learning_rate": 0.000633398166157266,
-      "loss": 0.1904,
+      "grad_norm": 0.6927788257598877,
+      "learning_rate": 0.0006333796424932852,
+      "loss": 0.1906,
       "step": 9000
     },
     {
       "epoch": 2.600722422895249,
-      "eval_loss": 0.19970637559890747,
-      "eval_runtime": 4.9095,
-      "eval_samples_per_second": 101.843,
-      "eval_steps_per_second": 6.518,
+      "eval_loss": 0.1991214007139206,
+      "eval_runtime": 4.7026,
+      "eval_samples_per_second": 106.324,
+      "eval_steps_per_second": 6.805,
       "step": 9360
     },
     {
       "epoch": 2.639622117254793,
-      "grad_norm": Infinity,
-      "learning_rate": 0.0006241363341668982,
-      "loss": 0.1928,
+      "grad_norm": 0.6956183910369873,
+      "learning_rate": 0.0006241178105029175,
+      "loss": 0.1893,
       "step": 9500
     },
     {
       "epoch": 2.7785495971103087,
-      "grad_norm": 0.5789324045181274,
-      "learning_rate": 0.0006148930258405113,
-      "loss": 0.1914,
+      "grad_norm": 0.40230458974838257,
+      "learning_rate": 0.0006148559785125498,
+      "loss": 0.1889,
       "step": 10000
     },
     {
       "epoch": 2.800777993887191,
-      "eval_loss": 0.20011071860790253,
-      "eval_runtime": 4.9182,
-      "eval_samples_per_second": 101.663,
-      "eval_steps_per_second": 6.506,
+      "eval_loss": 0.20080548524856567,
+      "eval_runtime": 4.5769,
+      "eval_samples_per_second": 109.245,
+      "eval_steps_per_second": 6.992,
       "step": 10080
     },
     {
       "epoch": 2.917477076965824,
-      "grad_norm": 0.5479323863983154,
-      "learning_rate": 0.0006056311938501436,
-      "loss": 0.193,
+      "grad_norm": 1.1422420740127563,
+      "learning_rate": 0.0006055941465221821,
+      "loss": 0.1911,
       "step": 10500
     },
     {
       "epoch": 3.000833564879133,
-      "eval_loss": 0.19800344109535217,
-      "eval_runtime": 5.4208,
-      "eval_samples_per_second": 92.236,
-      "eval_steps_per_second": 5.903,
+      "eval_loss": 0.2006545066833496,
+      "eval_runtime": 4.7461,
+      "eval_samples_per_second": 105.351,
+      "eval_steps_per_second": 6.742,
       "step": 10800
     },
     {
       "epoch": 3.0564045568213394,
-      "grad_norm": 0.509381890296936,
-      "learning_rate": 0.0005963693618597759,
-      "loss": 0.1849,
+      "grad_norm": 0.8359866142272949,
+      "learning_rate": 0.0005963323145318144,
+      "loss": 0.1858,
       "step": 11000
     },
     {
       "epoch": 3.1953320366768545,
-      "grad_norm": 0.5641015768051147,
-      "learning_rate": 0.0005871075298694081,
-      "loss": 0.1803,
+      "grad_norm": 0.5189564824104309,
+      "learning_rate": 0.0005870704825414468,
+      "loss": 0.1789,
       "step": 11500
     },
     {
       "epoch": 3.2008891358710754,
-      "eval_loss": 0.19798952341079712,
-      "eval_runtime": 6.0408,
-      "eval_samples_per_second": 82.771,
-      "eval_steps_per_second": 5.297,
+      "eval_loss": 0.19731611013412476,
+      "eval_runtime": 4.0407,
+      "eval_samples_per_second": 123.742,
+      "eval_steps_per_second": 7.919,
       "step": 11520
     },
     {
       "epoch": 3.33425951653237,
-      "grad_norm": 0.5248855352401733,
-      "learning_rate": 0.0005778456978790405,
-      "loss": 0.1785,
+      "grad_norm": 0.578628659248352,
+      "learning_rate": 0.0005778271742150597,
+      "loss": 0.1799,
       "step": 12000
     },
     {
       "epoch": 3.4009447068630174,
-      "eval_loss": 0.1982451230287552,
-      "eval_runtime": 6.0297,
-      "eval_samples_per_second": 82.922,
-      "eval_steps_per_second": 5.307,
+      "eval_loss": 0.19467230141162872,
+      "eval_runtime": 4.0481,
+      "eval_samples_per_second": 123.514,
+      "eval_steps_per_second": 7.905,
       "step": 12240
     },
     {
       "epoch": 3.4731869963878856,
-      "grad_norm": 1.105663537979126,
-      "learning_rate": 0.0005685838658886728,
-      "loss": 0.1758,
+      "grad_norm": 0.8292349576950073,
+      "learning_rate": 0.0005685653422246921,
+      "loss": 0.1769,
       "step": 12500
     },
     {
       "epoch": 3.6010002778549595,
-      "eval_loss": 0.1906416267156601,
-      "eval_runtime": 5.9927,
-      "eval_samples_per_second": 83.434,
-      "eval_steps_per_second": 5.34,
+      "eval_loss": 0.19331130385398865,
+      "eval_runtime": 4.0814,
+      "eval_samples_per_second": 122.508,
+      "eval_steps_per_second": 7.841,
       "step": 12960
     },
     {
       "epoch": 3.612114476243401,
-      "grad_norm": 1.019622802734375,
-      "learning_rate": 0.0005593405575622859,
-      "loss": 0.1765,
+      "grad_norm": 0.6857467293739319,
+      "learning_rate": 0.0005593035102343244,
+      "loss": 0.1796,
       "step": 13000
     },
     {
       "epoch": 3.7510419560989163,
-      "grad_norm": 0.8380174040794373,
-      "learning_rate": 0.0005500787255719182,
-      "loss": 0.1756,
+      "grad_norm": 0.5829126238822937,
+      "learning_rate": 0.0005500416782439567,
+      "loss": 0.176,
       "step": 13500
     },
     {
       "epoch": 3.801055848846902,
-      "eval_loss": 0.18713021278381348,
-      "eval_runtime": 4.8131,
-      "eval_samples_per_second": 103.884,
-      "eval_steps_per_second": 6.649,
+      "eval_loss": 0.18971021473407745,
+      "eval_runtime": 4.5985,
+      "eval_samples_per_second": 108.73,
+      "eval_steps_per_second": 6.959,
       "step": 13680
     },
     {
       "epoch": 3.889969435954432,
-      "grad_norm": 0.858310878276825,
-      "learning_rate": 0.0005408168935815504,
-      "loss": 0.1773,
+      "grad_norm": 0.7102944254875183,
+      "learning_rate": 0.0005407798462535889,
+      "loss": 0.1753,
       "step": 14000
     },
     {
       "epoch": 4.001111419838844,
-      "eval_loss": 0.18768689036369324,
-      "eval_runtime": 5.2594,
-      "eval_samples_per_second": 95.068,
-      "eval_steps_per_second": 6.084,
+      "eval_loss": 0.18332147598266602,
+      "eval_runtime": 4.569,
+      "eval_samples_per_second": 109.434,
+      "eval_steps_per_second": 7.004,
       "step": 14400
     },
     {
       "epoch": 4.0288969158099475,
-      "grad_norm": 0.8781201243400574,
-      "learning_rate": 0.0005315550615911827,
-      "loss": 0.1728,
+      "grad_norm": 0.5539807081222534,
+      "learning_rate": 0.000531536537927202,
+      "loss": 0.17,
       "step": 14500
     },
     {
       "epoch": 4.167824395665463,
-      "grad_norm": 0.6496415734291077,
-      "learning_rate": 0.0005223117532647958,
-      "loss": 0.1631,
+      "grad_norm": 0.6355165243148804,
+      "learning_rate": 0.0005222747059368344,
+      "loss": 0.1645,
       "step": 15000
     },
     {
       "epoch": 4.201166990830786,
-      "eval_loss": 0.18398858606815338,
-      "eval_runtime": 5.9345,
-      "eval_samples_per_second": 84.253,
-      "eval_steps_per_second": 5.392,
+      "eval_loss": 0.18329477310180664,
+      "eval_runtime": 4.6958,
+      "eval_samples_per_second": 106.477,
+      "eval_steps_per_second": 6.815,
       "step": 15120
     },
     {
       "epoch": 4.306751875520978,
-      "grad_norm": 0.6593993306159973,
-      "learning_rate": 0.0005130499212744282,
-      "loss": 0.1665,
+      "grad_norm": 0.5928105115890503,
+      "learning_rate": 0.0005130128739464667,
+      "loss": 0.1684,
       "step": 15500
     },
     {
       "epoch": 4.4012225618227285,
-      "eval_loss": 0.1805465966463089,
-      "eval_runtime": 6.0284,
-      "eval_samples_per_second": 82.941,
-      "eval_steps_per_second": 5.308,
+      "eval_loss": 0.1863529086112976,
+      "eval_runtime": 3.9814,
+      "eval_samples_per_second": 125.585,
+      "eval_steps_per_second": 8.037,
       "step": 15840
     },
     {
       "epoch": 4.445679355376494,
-      "grad_norm": 0.43227270245552063,
-      "learning_rate": 0.0005037880892840604,
-      "loss": 0.1658,
+      "grad_norm": 0.5573757886886597,
+      "learning_rate": 0.000503751041956099,
+      "loss": 0.1671,
       "step": 16000
     },
     {
       "epoch": 4.584606835232009,
-      "grad_norm": 0.6705788969993591,
-      "learning_rate": 0.0004945262572936927,
-      "loss": 0.1625,
+      "grad_norm": 0.5758992433547974,
+      "learning_rate": 0.0004944892099657312,
+      "loss": 0.1626,
       "step": 16500
     },
     {
       "epoch": 4.601278132814671,
-      "eval_loss": 0.18674401938915253,
-      "eval_runtime": 4.8422,
-      "eval_samples_per_second": 103.259,
-      "eval_steps_per_second": 6.609,
+      "eval_loss": 0.18393146991729736,
+      "eval_runtime": 3.9565,
+      "eval_samples_per_second": 126.373,
+      "eval_steps_per_second": 8.088,
       "step": 16560
     },
     {
       "epoch": 4.723534315087524,
-      "grad_norm": 0.9308199882507324,
-      "learning_rate": 0.0004852829489673058,
-      "loss": 0.164,
+      "grad_norm": 0.65135258436203,
+      "learning_rate": 0.0004852459016393443,
+      "loss": 0.1649,
       "step": 17000
     },
     {
       "epoch": 4.801333703806613,
-      "eval_loss": 0.17677545547485352,
-      "eval_runtime": 4.874,
-      "eval_samples_per_second": 102.584,
-      "eval_steps_per_second": 6.565,
+      "eval_loss": 0.176134392619133,
+      "eval_runtime": 4.7856,
+      "eval_samples_per_second": 104.479,
+      "eval_steps_per_second": 6.687,
       "step": 17280
     },
     {
       "epoch": 4.86246179494304,
-      "grad_norm": 0.488813579082489,
-      "learning_rate": 0.0004760211169769381,
-      "loss": 0.1645,
+      "grad_norm": 0.4339986741542816,
+      "learning_rate": 0.0004759840696489766,
+      "loss": 0.1643,
       "step": 17500
     },
     {
       "epoch": 5.001389274798555,
-      "grad_norm": 0.7509089708328247,
-      "learning_rate": 0.00046675928498657033,
-      "loss": 0.1593,
+      "grad_norm": 0.6604339480400085,
+      "learning_rate": 0.00046672223765860887,
+      "loss": 0.1613,
       "step": 18000
     },
     {
       "epoch": 5.001389274798555,
-      "eval_loss": 0.1795799285173416,
-      "eval_runtime": 6.143,
-      "eval_samples_per_second": 81.393,
-      "eval_steps_per_second": 5.209,
+      "eval_loss": 0.18032513558864594,
+      "eval_runtime": 4.4011,
+      "eval_samples_per_second": 113.608,
+      "eval_steps_per_second": 7.271,
       "step": 18000
     },
     {
       "epoch": 5.14031675465407,
-      "grad_norm": 0.8051738739013672,
-      "learning_rate": 0.0004574974529962027,
-      "loss": 0.1508,
+      "grad_norm": 0.5034669637680054,
+      "learning_rate": 0.00045746040566824116,
+      "loss": 0.1523,
       "step": 18500
     },
     {
       "epoch": 5.201444845790498,
-      "eval_loss": 0.1730128526687622,
-      "eval_runtime": 6.1102,
-      "eval_samples_per_second": 81.831,
-      "eval_steps_per_second": 5.237,
+      "eval_loss": 0.17549069225788116,
+      "eval_runtime": 4.1969,
+      "eval_samples_per_second": 119.137,
+      "eval_steps_per_second": 7.625,
       "step": 18720
     },
     {
       "epoch": 5.279244234509586,
-      "grad_norm": 0.616165816783905,
-      "learning_rate": 0.000448235621005835,
-      "loss": 0.151,
+      "grad_norm": 0.6701804995536804,
+      "learning_rate": 0.00044821709734185424,
+      "loss": 0.1531,
       "step": 19000
     },
     {
       "epoch": 5.401500416782439,
-      "eval_loss": 0.1723448783159256,
-      "eval_runtime": 4.9328,
-      "eval_samples_per_second": 101.362,
-      "eval_steps_per_second": 6.487,
+      "eval_loss": 0.1766393482685089,
+      "eval_runtime": 4.6747,
+      "eval_samples_per_second": 106.959,
+      "eval_steps_per_second": 6.845,
       "step": 19440
     },
     {
       "epoch": 5.418171714365101,
-      "grad_norm": 0.599356472492218,
-      "learning_rate": 0.00043897378901546733,
-      "loss": 0.152,
+      "grad_norm": 0.8808122277259827,
+      "learning_rate": 0.0004389552653514866,
+      "loss": 0.1538,
       "step": 19500
     },
     {
       "epoch": 5.5570991942206165,
-      "grad_norm": 0.6316538453102112,
-      "learning_rate": 0.0004297304806890803,
-      "loss": 0.1519,
+      "grad_norm": 0.7931129336357117,
+      "learning_rate": 0.0004296934333611189,
+      "loss": 0.1539,
       "step": 20000
     },
     {
       "epoch": 5.601555987774382,
-      "eval_loss": 0.17218750715255737,
-      "eval_runtime": 4.8261,
-      "eval_samples_per_second": 103.604,
-      "eval_steps_per_second": 6.631,
+      "eval_loss": 0.1726432889699936,
+      "eval_runtime": 3.9865,
+      "eval_samples_per_second": 125.423,
+      "eval_steps_per_second": 8.027,
       "step": 20160
     },
     {
       "epoch": 5.6960266740761325,
-      "grad_norm": 0.5698005557060242,
-      "learning_rate": 0.0004204686486987126,
-      "loss": 0.1522,
+      "grad_norm": 0.6453366875648499,
+      "learning_rate": 0.00042043160137075113,
+      "loss": 0.1527,
       "step": 20500
     },
     {
       "epoch": 5.801611558766324,
-      "eval_loss": 0.17050494253635406,
-      "eval_runtime": 6.0393,
-      "eval_samples_per_second": 82.791,
-      "eval_steps_per_second": 5.299,
+      "eval_loss": 0.17778918147087097,
+      "eval_runtime": 4.6178,
+      "eval_samples_per_second": 108.277,
+      "eval_steps_per_second": 6.93,
       "step": 20880
     },
     {
       "epoch": 5.834954153931648,
-      "grad_norm": 0.9362701773643494,
-      "learning_rate": 0.00041120681670834495,
-      "loss": 0.1497,
+      "grad_norm": 0.539543867111206,
+      "learning_rate": 0.0004111697693803834,
+      "loss": 0.1522,
       "step": 21000
     },
     {
       "epoch": 5.973881633787163,
-      "grad_norm": 0.7077497839927673,
-      "learning_rate": 0.00040194498471797724,
-      "loss": 0.1465,
+      "grad_norm": 0.7059842944145203,
+      "learning_rate": 0.0004019079373900158,
+      "loss": 0.1483,
       "step": 21500
     },
     {
       "epoch": 6.001667129758266,
-      "eval_loss": 0.1704930067062378,
-      "eval_runtime": 6.2406,
-      "eval_samples_per_second": 80.12,
-      "eval_steps_per_second": 5.128,
+      "eval_loss": 0.16399218142032623,
+      "eval_runtime": 4.5977,
+      "eval_samples_per_second": 108.749,
+      "eval_steps_per_second": 6.96,
       "step": 21600
     },
     {
       "epoch": 6.112809113642679,
-      "grad_norm": 0.6819673180580139,
-      "learning_rate": 0.00039268315272760954,
-      "loss": 0.1391,
+      "grad_norm": 0.6448105573654175,
+      "learning_rate": 0.00039264610539964807,
+      "loss": 0.1404,
       "step": 22000
     },
     {
       "epoch": 6.201722700750208,
-      "eval_loss": 0.1641210913658142,
-      "eval_runtime": 6.0676,
-      "eval_samples_per_second": 82.405,
-      "eval_steps_per_second": 5.274,
+      "eval_loss": 0.16838781535625458,
+      "eval_runtime": 4.653,
+      "eval_samples_per_second": 107.459,
+      "eval_steps_per_second": 6.877,
       "step": 22320
     },
     {
       "epoch": 6.251736593498194,
-      "grad_norm": 0.6154927015304565,
-      "learning_rate": 0.0003834213207372419,
-      "loss": 0.1383,
+      "grad_norm": 0.6864171028137207,
+      "learning_rate": 0.0003834027970732611,
+      "loss": 0.1425,
       "step": 22500
     },
     {
       "epoch": 6.390664073353709,
-      "grad_norm": 0.5655333399772644,
-      "learning_rate": 0.00037415948874687413,
-      "loss": 0.1392,
+      "grad_norm": 0.6795888543128967,
+      "learning_rate": 0.0003741409650828934,
+      "loss": 0.1426,
       "step": 23000
     },
     {
       "epoch": 6.401778271742151,
-      "eval_loss": 0.16602876782417297,
-      "eval_runtime": 4.8554,
-      "eval_samples_per_second": 102.978,
-      "eval_steps_per_second": 6.591,
+      "eval_loss": 0.1665893793106079,
+      "eval_runtime": 4.626,
+      "eval_samples_per_second": 108.085,
+      "eval_steps_per_second": 6.917,
       "step": 23040
     },
     {
       "epoch": 6.529591553209225,
-      "grad_norm": 0.6170830726623535,
-      "learning_rate": 0.0003648976567565065,
-      "loss": 0.1374,
+      "grad_norm": 0.6111213564872742,
+      "learning_rate": 0.00036487913309252574,
+      "loss": 0.1407,
       "step": 23500
     },
     {
       "epoch": 6.601833842734093,
-      "eval_loss": 0.16095949709415436,
-      "eval_runtime": 4.8435,
-      "eval_samples_per_second": 103.232,
-      "eval_steps_per_second": 6.607,
+      "eval_loss": 0.16504846513271332,
+      "eval_runtime": 3.8884,
+      "eval_samples_per_second": 128.587,
+      "eval_steps_per_second": 8.23,
       "step": 23760
     },
     {
       "epoch": 6.66851903306474,
-      "grad_norm": 0.7002782821655273,
-      "learning_rate": 0.0003556543484301195,
-      "loss": 0.1369,
+      "grad_norm": 0.5933897495269775,
+      "learning_rate": 0.000355617301102158,
+      "loss": 0.1413,
       "step": 24000
     },
     {
       "epoch": 6.801889413726035,
-      "eval_loss": 0.1629977971315384,
-      "eval_runtime": 6.0015,
-      "eval_samples_per_second": 83.313,
-      "eval_steps_per_second": 5.332,
+      "eval_loss": 0.16348223388195038,
+      "eval_runtime": 4.6832,
+      "eval_samples_per_second": 106.764,
+      "eval_steps_per_second": 6.833,
       "step": 24480
     },
     {
       "epoch": 6.807446512920255,
-      "grad_norm": 0.6815934777259827,
-      "learning_rate": 0.0003463925164397518,
-      "loss": 0.1372,
+      "grad_norm": 0.7584686875343323,
+      "learning_rate": 0.00034635546911179033,
+      "loss": 0.1405,
       "step": 24500
     },
     {
       "epoch": 6.946373992775771,
-      "grad_norm": 0.6028130650520325,
-      "learning_rate": 0.0003371306844493841,
-      "loss": 0.1358,
+      "grad_norm": 0.614815890789032,
+      "learning_rate": 0.00033709363712142263,
+      "loss": 0.1386,
       "step": 25000
     },
     {
       "epoch": 7.001944984717977,
-      "eval_loss": 0.15650975704193115,
-      "eval_runtime": 6.1328,
-      "eval_samples_per_second": 81.529,
-      "eval_steps_per_second": 5.218,
+      "eval_loss": 0.16397513449192047,
+      "eval_runtime": 4.5957,
+      "eval_samples_per_second": 108.798,
+      "eval_steps_per_second": 6.963,
       "step": 25200
     },
     {
       "epoch": 7.085301472631286,
-      "grad_norm": 0.8667997717857361,
-      "learning_rate": 0.0003278873761229972,
-      "loss": 0.1258,
+      "grad_norm": 0.5844869613647461,
+      "learning_rate": 0.0003278318051310549,
+      "loss": 0.1307,
       "step": 25500
     },
     {
       "epoch": 7.20200055570992,
-      "eval_loss": 0.16240185499191284,
-      "eval_runtime": 6.0855,
-      "eval_samples_per_second": 82.162,
-      "eval_steps_per_second": 5.258,
+      "eval_loss": 0.1627485603094101,
+      "eval_runtime": 3.892,
+      "eval_samples_per_second": 128.47,
+      "eval_steps_per_second": 8.222,
       "step": 25920
     },
     {
       "epoch": 7.2242289524868015,
-      "grad_norm": 0.7652863264083862,
-      "learning_rate": 0.0003186255441326294,
-      "loss": 0.1247,
+      "grad_norm": 0.5960623025894165,
+      "learning_rate": 0.0003185699731406873,
+      "loss": 0.1282,
       "step": 26000
     },
     {
       "epoch": 7.3631564323423175,
-      "grad_norm": 1.3608860969543457,
-      "learning_rate": 0.00030936371214226177,
-      "loss": 0.1259,
+      "grad_norm": 0.5819278359413147,
+      "learning_rate": 0.0003093081411503195,
+      "loss": 0.1291,
       "step": 26500
     },
     {
       "epoch": 7.402056126701861,
-      "eval_loss": 0.1604655385017395,
-      "eval_runtime": 6.0515,
-      "eval_samples_per_second": 82.624,
-      "eval_steps_per_second": 5.288,
+      "eval_loss": 0.16200314462184906,
+      "eval_runtime": 4.5372,
+      "eval_samples_per_second": 110.199,
+      "eval_steps_per_second": 7.053,
       "step": 26640
     },
     {
       "epoch": 7.502083912197833,
-      "grad_norm": 0.48411527276039124,
-      "learning_rate": 0.00030010188015189407,
-      "loss": 0.1268,
+      "grad_norm": 0.4887467920780182,
+      "learning_rate": 0.0003000648328239326,
+      "loss": 0.1293,
       "step": 27000
     },
     {
       "epoch": 7.602111697693804,
-      "eval_loss": 0.15553823113441467,
-      "eval_runtime": 4.8782,
-      "eval_samples_per_second": 102.497,
-      "eval_steps_per_second": 6.56,
+      "eval_loss": 0.16259238123893738,
+      "eval_runtime": 4.6942,
+      "eval_samples_per_second": 106.513,
+      "eval_steps_per_second": 6.817,
       "step": 27360
     },
     {
       "epoch": 7.641011392053348,
-      "grad_norm": 0.6623326539993286,
-      "learning_rate": 0.0002908585718255071,
-      "loss": 0.1247,
+      "grad_norm": 0.43387722969055176,
+      "learning_rate": 0.0002908030008335649,
+      "loss": 0.1291,
       "step": 27500
     },
     {
       "epoch": 7.779938871908864,
-      "grad_norm": 0.5693512558937073,
-      "learning_rate": 0.0002815967398351394,
-      "loss": 0.1231,
+      "grad_norm": 0.6429397463798523,
+      "learning_rate": 0.0002815411688431972,
+      "loss": 0.128,
       "step": 28000
     },
     {
       "epoch": 7.802167268685746,
-      "eval_loss": 0.15078049898147583,
-      "eval_runtime": 5.5749,
-      "eval_samples_per_second": 89.687,
-      "eval_steps_per_second": 5.74,
+      "eval_loss": 0.15840420126914978,
+      "eval_runtime": 3.9278,
+      "eval_samples_per_second": 127.299,
+      "eval_steps_per_second": 8.147,
       "step": 28080
     },
     {
       "epoch": 7.918866351764379,
-      "grad_norm": 0.5414377450942993,
-      "learning_rate": 0.0002723349078447717,
-      "loss": 0.1244,
+      "grad_norm": 0.47245821356773376,
+      "learning_rate": 0.0002722978605168102,
+      "loss": 0.1288,
       "step": 28500
     },
     {
       "epoch": 8.002222839677689,
-      "eval_loss": 0.1523188203573227,
-      "eval_runtime": 6.1165,
-      "eval_samples_per_second": 81.746,
-      "eval_steps_per_second": 5.232,
+      "eval_loss": 0.1579759120941162,
+      "eval_runtime": 4.1778,
+      "eval_samples_per_second": 119.68,
+      "eval_steps_per_second": 7.66,
       "step": 28800
     },
     {
       "epoch": 8.057793831619895,
-      "grad_norm": 0.7959994077682495,
-      "learning_rate": 0.000263073075854404,
-      "loss": 0.1177,
+      "grad_norm": 0.6740341782569885,
+      "learning_rate": 0.00026303602852644257,
+      "loss": 0.1217,
       "step": 29000
     },
     {
       "epoch": 8.19672131147541,
-      "grad_norm": 0.7818427085876465,
-      "learning_rate": 0.00025381124386403633,
-      "loss": 0.1125,
+      "grad_norm": 0.6444827318191528,
+      "learning_rate": 0.00025377419653607486,
+      "loss": 0.1182,
       "step": 29500
     },
     {
       "epoch": 8.20227841066963,
-      "eval_loss": 0.1530105322599411,
-      "eval_runtime": 5.9895,
-      "eval_samples_per_second": 83.479,
-      "eval_steps_per_second": 5.343,
+      "eval_loss": 0.15706072747707367,
+      "eval_runtime": 4.5983,
+      "eval_samples_per_second": 108.735,
+      "eval_steps_per_second": 6.959,
       "step": 29520
     },
     {
       "epoch": 8.335648791330925,
-      "grad_norm": 0.6304500699043274,
-      "learning_rate": 0.00024456793553764936,
-      "loss": 0.1107,
+      "grad_norm": 0.6599079370498657,
+      "learning_rate": 0.00024451236454570716,
+      "loss": 0.1166,
       "step": 30000
     },
     {
       "epoch": 8.402333981661572,
-      "eval_loss": 0.15071706473827362,
-      "eval_runtime": 6.0586,
-      "eval_samples_per_second": 82.528,
-      "eval_steps_per_second": 5.282,
+      "eval_loss": 0.15387499332427979,
+      "eval_runtime": 4.1869,
+      "eval_samples_per_second": 119.419,
+      "eval_steps_per_second": 7.643,
       "step": 30240
     },
     {
       "epoch": 8.474576271186441,
-      "grad_norm": 0.5695451498031616,
-      "learning_rate": 0.00023530610354728166,
-      "loss": 0.1125,
+      "grad_norm": 0.4398277997970581,
+      "learning_rate": 0.00023525053255533945,
+      "loss": 0.1178,
       "step": 30500
     },
     {
       "epoch": 8.602389552653515,
-      "eval_loss": 0.15306152403354645,
-      "eval_runtime": 5.0581,
-      "eval_samples_per_second": 98.851,
-      "eval_steps_per_second": 6.326,
+      "eval_loss": 0.15714193880558014,
+      "eval_runtime": 3.9157,
+      "eval_samples_per_second": 127.69,
+      "eval_steps_per_second": 8.172,
       "step": 30960
     },
     {
       "epoch": 8.613503751041955,
-      "grad_norm": 0.6288810968399048,
-      "learning_rate": 0.00022604427155691398,
-      "loss": 0.1128,
+      "grad_norm": 0.7029595971107483,
+      "learning_rate": 0.00022598870056497178,
+      "loss": 0.1188,
       "step": 31000
     },
     {
       "epoch": 8.752431230897471,
-      "grad_norm": 0.5629562139511108,
-      "learning_rate": 0.00021678243956654627,
-      "loss": 0.1109,
+      "grad_norm": 0.8394140601158142,
+      "learning_rate": 0.00021672686857460404,
+      "loss": 0.1154,
       "step": 31500
     },
     {
       "epoch": 8.802445123645457,
-      "eval_loss": 0.14975672960281372,
-      "eval_runtime": 6.0264,
-      "eval_samples_per_second": 82.968,
-      "eval_steps_per_second": 5.31,
+      "eval_loss": 0.15471811592578888,
+      "eval_runtime": 4.5747,
+      "eval_samples_per_second": 109.297,
+      "eval_steps_per_second": 6.995,
       "step": 31680
     },
     {
       "epoch": 8.891358710752987,
-      "grad_norm": 0.7416991591453552,
-      "learning_rate": 0.0002075206075761786,
-      "loss": 0.1115,
+      "grad_norm": 0.5649506449699402,
+      "learning_rate": 0.0002074835602482171,
+      "loss": 0.1163,
       "step": 32000
     }
   ],

checkpoint-32391/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:43d31253d1d9828754b954b23155d823529e2ebc7f148e5c229e9bc7b696124d
-size 6033
+oid sha256:1ac71e3cdd10165b76004d5b1c4b8cd4a183b94ec338868fce8dc9960bf7c890
+size 5496