Training in progress, step 30, checkpoint

checkpoint-30/adapter_config.json

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

checkpoint-30/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a1a9e0b590ffdc0fac6d4895f608ffd2d0753ab395a71921639c2f9fd61ca545
+oid sha256:e8b97773e04827e934022820bfa7fbd5bad93af3e2ed70e773e85740afe2cc6c
 size 80013120

checkpoint-30/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9e7d2f51158926f839deaf2a2cb5ca95934615b5a99cad711ecdb2aab1e8f45f
+oid sha256:476729e13139d92a235311e1b8c488d8270f6b819b7960034a803a04f45d5278
 size 160284754

checkpoint-30/trainer_state.json

@@ -10,292 +10,292 @@
   "log_history": [
     {
       "epoch": 0.0006200589055960316,
-      "grad_norm": 0.18599237501621246,
+      "grad_norm": 0.16878783702850342,
       "learning_rate": 4.000000000000001e-06,
-      "loss": 0.9319,
+      "loss": 0.9729,
       "step": 1
     },
     {
       "epoch": 0.0012401178111920632,
-      "grad_norm": 0.09086523950099945,
+      "grad_norm": 0.0865061804652214,
       "learning_rate": 8.000000000000001e-06,
-      "loss": 0.6463,
+      "loss": 0.6661,
       "step": 2
     },
     {
       "epoch": 0.0018601767167880949,
-      "grad_norm": 0.09325356036424637,
+      "grad_norm": 0.0920981764793396,
       "learning_rate": 1.2e-05,
-      "loss": 0.575,
+      "loss": 0.5954,
       "step": 3
     },
     {
       "epoch": 0.0018601767167880949,
-      "eval_loss": 0.7272388935089111,
-      "eval_runtime": 33.9288,
-      "eval_samples_per_second": 2.947,
-      "eval_steps_per_second": 2.947,
+      "eval_loss": 0.762992262840271,
+      "eval_runtime": 33.3972,
+      "eval_samples_per_second": 2.994,
+      "eval_steps_per_second": 2.994,
       "step": 3
     },
     {
       "epoch": 0.0024802356223841263,
-      "grad_norm": 0.09707680344581604,
+      "grad_norm": 0.10451442003250122,
       "learning_rate": 1.6000000000000003e-05,
-      "loss": 0.7038,
+      "loss": 0.7227,
       "step": 4
     },
     {
       "epoch": 0.0031002945279801583,
-      "grad_norm": 0.10810782760381699,
+      "grad_norm": 0.11351602524518967,
       "learning_rate": 2e-05,
-      "loss": 0.5471,
+      "loss": 0.5719,
       "step": 5
     },
     {
       "epoch": 0.0037203534335761897,
-      "grad_norm": 0.13030411303043365,
+      "grad_norm": 0.12484906613826752,
       "learning_rate": 1.9200000000000003e-05,
-      "loss": 0.6735,
+      "loss": 0.6972,
       "step": 6
     },
     {
       "epoch": 0.0037203534335761897,
-      "eval_loss": 0.7256982922554016,
-      "eval_runtime": 33.9528,
-      "eval_samples_per_second": 2.945,
-      "eval_steps_per_second": 2.945,
+      "eval_loss": 0.7615777850151062,
+      "eval_runtime": 33.3804,
+      "eval_samples_per_second": 2.996,
+      "eval_steps_per_second": 2.996,
       "step": 6
     },
     {
       "epoch": 0.004340412339172222,
-      "grad_norm": 0.12336976826190948,
+      "grad_norm": 0.11116621643304825,
       "learning_rate": 1.8400000000000003e-05,
-      "loss": 0.654,
+      "loss": 0.6789,
       "step": 7
     },
     {
       "epoch": 0.004960471244768253,
-      "grad_norm": 0.16046489775180817,
+      "grad_norm": 0.16013608872890472,
       "learning_rate": 1.76e-05,
-      "loss": 0.6301,
+      "loss": 0.6612,
       "step": 8
     },
     {
       "epoch": 0.005580530150364285,
-      "grad_norm": 0.11589628458023071,
+      "grad_norm": 0.1211152970790863,
       "learning_rate": 1.6800000000000002e-05,
-      "loss": 0.5326,
+      "loss": 0.5557,
       "step": 9
     },
     {
       "epoch": 0.005580530150364285,
-      "eval_loss": 0.723076343536377,
-      "eval_runtime": 34.5303,
-      "eval_samples_per_second": 2.896,
-      "eval_steps_per_second": 2.896,
+      "eval_loss": 0.7586225867271423,
+      "eval_runtime": 33.7015,
+      "eval_samples_per_second": 2.967,
+      "eval_steps_per_second": 2.967,
       "step": 9
     },
     {
       "epoch": 0.0062005890559603165,
-      "grad_norm": 0.12244553864002228,
+      "grad_norm": 0.1299518495798111,
       "learning_rate": 1.6000000000000003e-05,
-      "loss": 0.8055,
+      "loss": 0.8339,
       "step": 10
     },
     {
       "epoch": 0.0068206479615563476,
-      "grad_norm": 0.10270956158638,
+      "grad_norm": 0.11476285755634308,
       "learning_rate": 1.5200000000000002e-05,
-      "loss": 0.5492,
+      "loss": 0.5786,
       "step": 11
     },
     {
       "epoch": 0.0074407068671523795,
-      "grad_norm": 0.14439155161380768,
+      "grad_norm": 0.18253028392791748,
       "learning_rate": 1.4400000000000001e-05,
-      "loss": 0.8579,
+      "loss": 0.8889,
       "step": 12
     },
     {
       "epoch": 0.0074407068671523795,
-      "eval_loss": 0.7205619215965271,
-      "eval_runtime": 33.7297,
-      "eval_samples_per_second": 2.965,
-      "eval_steps_per_second": 2.965,
+      "eval_loss": 0.7548640966415405,
+      "eval_runtime": 34.763,
+      "eval_samples_per_second": 2.877,
+      "eval_steps_per_second": 2.877,
       "step": 12
     },
     {
       "epoch": 0.008060765772748411,
-      "grad_norm": 0.1271936148405075,
+      "grad_norm": 0.14364618062973022,
       "learning_rate": 1.3600000000000002e-05,
-      "loss": 0.6923,
+      "loss": 0.7251,
       "step": 13
     },
     {
       "epoch": 0.008680824678344443,
-      "grad_norm": 0.092640720307827,
+      "grad_norm": 0.11001864820718765,
       "learning_rate": 1.2800000000000001e-05,
-      "loss": 0.5983,
+      "loss": 0.6209,
       "step": 14
     },
     {
       "epoch": 0.009300883583940475,
-      "grad_norm": 0.1161455512046814,
+      "grad_norm": 0.1239691898226738,
       "learning_rate": 1.2e-05,
-      "loss": 0.6386,
+      "loss": 0.6641,
       "step": 15
     },
     {
       "epoch": 0.009300883583940475,
-      "eval_loss": 0.7178443074226379,
-      "eval_runtime": 33.5878,
-      "eval_samples_per_second": 2.977,
-      "eval_steps_per_second": 2.977,
+      "eval_loss": 0.7505818009376526,
+      "eval_runtime": 34.0273,
+      "eval_samples_per_second": 2.939,
+      "eval_steps_per_second": 2.939,
       "step": 15
     },
     {
       "epoch": 0.009920942489536505,
-      "grad_norm": 0.09220079332590103,
+      "grad_norm": 0.11025531589984894,
       "learning_rate": 1.1200000000000001e-05,
-      "loss": 0.4665,
+      "loss": 0.4851,
       "step": 16
     },
     {
       "epoch": 0.010541001395132537,
-      "grad_norm": 0.08713295310735703,
+      "grad_norm": 0.10391498357057571,
       "learning_rate": 1.04e-05,
-      "loss": 0.6114,
+      "loss": 0.6274,
       "step": 17
     },
     {
       "epoch": 0.01116106030072857,
-      "grad_norm": 0.11913106590509415,
+      "grad_norm": 0.13879302144050598,
       "learning_rate": 9.600000000000001e-06,
-      "loss": 0.5703,
+      "loss": 0.5931,
       "step": 18
     },
     {
       "epoch": 0.01116106030072857,
-      "eval_loss": 0.7152560949325562,
-      "eval_runtime": 33.4806,
-      "eval_samples_per_second": 2.987,
-      "eval_steps_per_second": 2.987,
+      "eval_loss": 0.7464621067047119,
+      "eval_runtime": 34.3301,
+      "eval_samples_per_second": 2.913,
+      "eval_steps_per_second": 2.913,
       "step": 18
     },
     {
       "epoch": 0.011781119206324601,
-      "grad_norm": 0.11837327480316162,
+      "grad_norm": 0.14016444981098175,
       "learning_rate": 8.8e-06,
-      "loss": 0.6049,
+      "loss": 0.6289,
       "step": 19
     },
     {
       "epoch": 0.012401178111920633,
-      "grad_norm": 0.08001928776502609,
+      "grad_norm": 0.11601278930902481,
       "learning_rate": 8.000000000000001e-06,
-      "loss": 0.4691,
+      "loss": 0.4875,
       "step": 20
     },
     {
       "epoch": 0.013021237017516665,
-      "grad_norm": 0.11291633546352386,
+      "grad_norm": 0.13207638263702393,
       "learning_rate": 7.2000000000000005e-06,
-      "loss": 0.6752,
+      "loss": 0.6942,
       "step": 21
     },
     {
       "epoch": 0.013021237017516665,
-      "eval_loss": 0.7128345370292664,
-      "eval_runtime": 34.2622,
-      "eval_samples_per_second": 2.919,
-      "eval_steps_per_second": 2.919,
+      "eval_loss": 0.7427210807800293,
+      "eval_runtime": 34.6202,
+      "eval_samples_per_second": 2.888,
+      "eval_steps_per_second": 2.888,
       "step": 21
     },
     {
       "epoch": 0.013641295923112695,
-      "grad_norm": 0.1115325391292572,
+      "grad_norm": 0.12952734529972076,
       "learning_rate": 6.4000000000000006e-06,
-      "loss": 0.5874,
+      "loss": 0.6095,
       "step": 22
     },
     {
       "epoch": 0.014261354828708727,
-      "grad_norm": 0.1305752545595169,
+      "grad_norm": 0.15325698256492615,
       "learning_rate": 5.600000000000001e-06,
-      "loss": 0.7733,
+      "loss": 0.798,
       "step": 23
     },
     {
       "epoch": 0.014881413734304759,
-      "grad_norm": 0.10779058188199997,
+      "grad_norm": 0.12156686931848526,
       "learning_rate": 4.800000000000001e-06,
-      "loss": 0.6217,
+      "loss": 0.6426,
       "step": 24
     },
     {
       "epoch": 0.014881413734304759,
-      "eval_loss": 0.7108609080314636,
-      "eval_runtime": 34.6034,
-      "eval_samples_per_second": 2.89,
-      "eval_steps_per_second": 2.89,
+      "eval_loss": 0.7396314740180969,
+      "eval_runtime": 34.8471,
+      "eval_samples_per_second": 2.87,
+      "eval_steps_per_second": 2.87,
       "step": 24
     },
     {
       "epoch": 0.01550147263990079,
-      "grad_norm": 0.11457020789384842,
+      "grad_norm": 0.14465734362602234,
       "learning_rate": 4.000000000000001e-06,
-      "loss": 0.5838,
+      "loss": 0.6037,
       "step": 25
     },
     {
       "epoch": 0.016121531545496823,
-      "grad_norm": 0.1093108132481575,
+      "grad_norm": 0.12176154553890228,
       "learning_rate": 3.2000000000000003e-06,
-      "loss": 0.504,
+      "loss": 0.5222,
       "step": 26
     },
     {
       "epoch": 0.016741590451092855,
-      "grad_norm": 0.11828811466693878,
+      "grad_norm": 0.13855800032615662,
       "learning_rate": 2.4000000000000003e-06,
-      "loss": 0.5179,
+      "loss": 0.5386,
       "step": 27
     },
     {
       "epoch": 0.016741590451092855,
-      "eval_loss": 0.7094972133636475,
-      "eval_runtime": 34.2275,
-      "eval_samples_per_second": 2.922,
-      "eval_steps_per_second": 2.922,
+      "eval_loss": 0.7376267910003662,
+      "eval_runtime": 34.7753,
+      "eval_samples_per_second": 2.876,
+      "eval_steps_per_second": 2.876,
       "step": 27
     },
     {
       "epoch": 0.017361649356688887,
-      "grad_norm": 0.09534204006195068,
+      "grad_norm": 0.1102200523018837,
       "learning_rate": 1.6000000000000001e-06,
-      "loss": 0.4988,
+      "loss": 0.5182,
       "step": 28
     },
     {
       "epoch": 0.01798170826228492,
-      "grad_norm": 0.12317203730344772,
+      "grad_norm": 0.1387104094028473,
       "learning_rate": 8.000000000000001e-07,
-      "loss": 0.6945,
+      "loss": 0.716,
       "step": 29
     },
     {
       "epoch": 0.01860176716788095,
-      "grad_norm": 0.13696961104869843,
+      "grad_norm": 0.16438902914524078,
       "learning_rate": 0.0,
-      "loss": 0.5378,
+      "loss": 0.5635,
       "step": 30
     },
     {
       "epoch": 0.01860176716788095,
-      "eval_loss": 0.7089178562164307,
-      "eval_runtime": 33.3831,
-      "eval_samples_per_second": 2.996,
-      "eval_steps_per_second": 2.996,
+      "eval_loss": 0.7367461323738098,
+      "eval_runtime": 34.8255,
+      "eval_samples_per_second": 2.871,
+      "eval_steps_per_second": 2.871,
       "step": 30
     }
   ],
@@ -316,7 +316,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 4.029866007950131e+16,
+  "total_flos": 4.068039818310451e+16,
   "train_batch_size": 2,
   "trial_name": null,
   "trial_params": null

checkpoint-30/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bd165a72088a81e5441087b0ae01dd72f7f14c5dad6a2e407eebc961586b0fc2
+oid sha256:1fba20505ee5f941c5085404aed8036910f9919d2bc7b2ba5ba4186849388b0c
 size 5752