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:e42ff132ae9e27b25c14b1a0d2e98b623f5d8a78aac496813de619136d7fdb27
+oid sha256:6e73ad3aca4ce497d080cb8a6c619cd7f5891102a2eedc87026fe96b9db237ea
 size 131146352

checkpoint-30/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5a6f5950021d2a9b0bba0756d546bc9a2f40e210c0800cc0a3dfc50b8468c8e8
+oid sha256:f2ee9e574a683ed57cea4fc36a2dabd24fd7170daaab96491ae55ca3db567a46
 size 262615650

checkpoint-30/trainer_state.json

@@ -10,292 +10,292 @@
   "log_history": [
     {
       "epoch": 0.0006200589055960316,
-      "grad_norm": 0.39027148485183716,
+      "grad_norm": 0.5875879526138306,
       "learning_rate": 4.000000000000001e-06,
-      "loss": 1.1401,
+      "loss": 1.179,
       "step": 1
     },
     {
       "epoch": 0.0012401178111920632,
-      "grad_norm": 0.1772935390472412,
+      "grad_norm": 0.2559925317764282,
       "learning_rate": 8.000000000000001e-06,
-      "loss": 0.9133,
+      "loss": 0.9296,
       "step": 2
     },
     {
       "epoch": 0.0018601767167880949,
-      "grad_norm": 0.192855104804039,
+      "grad_norm": 0.27975356578826904,
       "learning_rate": 1.2e-05,
-      "loss": 0.8036,
+      "loss": 0.8212,
       "step": 3
     },
     {
       "epoch": 0.0018601767167880949,
-      "eval_loss": 0.993510365486145,
-      "eval_runtime": 46.5106,
-      "eval_samples_per_second": 2.15,
-      "eval_steps_per_second": 2.15,
+      "eval_loss": 1.0227885246276855,
+      "eval_runtime": 47.9985,
+      "eval_samples_per_second": 2.083,
+      "eval_steps_per_second": 2.083,
       "step": 3
     },
     {
       "epoch": 0.0024802356223841263,
-      "grad_norm": 0.23705194890499115,
+      "grad_norm": 0.4077853858470917,
       "learning_rate": 1.6000000000000003e-05,
-      "loss": 0.9389,
+      "loss": 0.9553,
       "step": 4
     },
     {
       "epoch": 0.0031002945279801583,
-      "grad_norm": 0.22702816128730774,
+      "grad_norm": 0.3481099307537079,
       "learning_rate": 2e-05,
-      "loss": 0.7836,
+      "loss": 0.8071,
       "step": 5
     },
     {
       "epoch": 0.0037203534335761897,
-      "grad_norm": 0.22573529183864594,
+      "grad_norm": 0.33538541197776794,
       "learning_rate": 1.9200000000000003e-05,
-      "loss": 0.8874,
+      "loss": 0.9046,
       "step": 6
     },
     {
       "epoch": 0.0037203534335761897,
-      "eval_loss": 0.9887735843658447,
-      "eval_runtime": 47.2769,
-      "eval_samples_per_second": 2.115,
-      "eval_steps_per_second": 2.115,
+      "eval_loss": 1.0122885704040527,
+      "eval_runtime": 48.1202,
+      "eval_samples_per_second": 2.078,
+      "eval_steps_per_second": 2.078,
       "step": 6
     },
     {
       "epoch": 0.004340412339172222,
-      "grad_norm": 0.231398344039917,
+      "grad_norm": 0.2848486602306366,
       "learning_rate": 1.8400000000000003e-05,
-      "loss": 0.8755,
+      "loss": 0.8963,
       "step": 7
     },
     {
       "epoch": 0.004960471244768253,
-      "grad_norm": 0.28229379653930664,
+      "grad_norm": 0.29836800694465637,
       "learning_rate": 1.76e-05,
-      "loss": 0.971,
+      "loss": 0.9869,
       "step": 8
     },
     {
       "epoch": 0.005580530150364285,
-      "grad_norm": 0.21991951763629913,
+      "grad_norm": 0.22438423335552216,
       "learning_rate": 1.6800000000000002e-05,
-      "loss": 0.8546,
+      "loss": 0.8704,
       "step": 9
     },
     {
       "epoch": 0.005580530150364285,
-      "eval_loss": 0.9774181842803955,
-      "eval_runtime": 47.3982,
-      "eval_samples_per_second": 2.11,
-      "eval_steps_per_second": 2.11,
+      "eval_loss": 0.996995210647583,
+      "eval_runtime": 48.1511,
+      "eval_samples_per_second": 2.077,
+      "eval_steps_per_second": 2.077,
       "step": 9
     },
     {
       "epoch": 0.0062005890559603165,
-      "grad_norm": 0.26028916239738464,
+      "grad_norm": 0.26462429761886597,
       "learning_rate": 1.6000000000000003e-05,
-      "loss": 1.0344,
+      "loss": 1.0479,
       "step": 10
     },
     {
       "epoch": 0.0068206479615563476,
-      "grad_norm": 0.23867206275463104,
+      "grad_norm": 0.2650766968727112,
       "learning_rate": 1.5200000000000002e-05,
-      "loss": 0.8132,
+      "loss": 0.8352,
       "step": 11
     },
     {
       "epoch": 0.0074407068671523795,
-      "grad_norm": 0.30641764402389526,
+      "grad_norm": 0.3595620095729828,
       "learning_rate": 1.4400000000000001e-05,
-      "loss": 1.1168,
+      "loss": 1.1297,
       "step": 12
     },
     {
       "epoch": 0.0074407068671523795,
-      "eval_loss": 0.963714599609375,
-      "eval_runtime": 47.188,
-      "eval_samples_per_second": 2.119,
-      "eval_steps_per_second": 2.119,
+      "eval_loss": 0.9821679592132568,
+      "eval_runtime": 48.205,
+      "eval_samples_per_second": 2.074,
+      "eval_steps_per_second": 2.074,
       "step": 12
     },
     {
       "epoch": 0.008060765772748411,
-      "grad_norm": 0.25963348150253296,
+      "grad_norm": 0.2854643166065216,
       "learning_rate": 1.3600000000000002e-05,
-      "loss": 0.9735,
+      "loss": 0.9925,
       "step": 13
     },
     {
       "epoch": 0.008680824678344443,
-      "grad_norm": 0.2063651829957962,
+      "grad_norm": 0.21441830694675446,
       "learning_rate": 1.2800000000000001e-05,
-      "loss": 0.7957,
+      "loss": 0.8125,
       "step": 14
     },
     {
       "epoch": 0.009300883583940475,
-      "grad_norm": 0.2287115454673767,
+      "grad_norm": 0.21920695900917053,
       "learning_rate": 1.2e-05,
-      "loss": 0.8343,
+      "loss": 0.8518,
       "step": 15
     },
     {
       "epoch": 0.009300883583940475,
-      "eval_loss": 0.9510444402694702,
-      "eval_runtime": 47.3972,
-      "eval_samples_per_second": 2.11,
-      "eval_steps_per_second": 2.11,
+      "eval_loss": 0.9693244695663452,
+      "eval_runtime": 47.8177,
+      "eval_samples_per_second": 2.091,
+      "eval_steps_per_second": 2.091,
       "step": 15
     },
     {
       "epoch": 0.009920942489536505,
-      "grad_norm": 0.2136881947517395,
+      "grad_norm": 0.20094920694828033,
       "learning_rate": 1.1200000000000001e-05,
-      "loss": 0.7861,
+      "loss": 0.8003,
       "step": 16
     },
     {
       "epoch": 0.010541001395132537,
-      "grad_norm": 0.16134004294872284,
+      "grad_norm": 0.21595342457294464,
       "learning_rate": 1.04e-05,
-      "loss": 0.8375,
+      "loss": 0.8492,
       "step": 17
     },
     {
       "epoch": 0.01116106030072857,
-      "grad_norm": 0.20020240545272827,
+      "grad_norm": 0.20084670186042786,
       "learning_rate": 9.600000000000001e-06,
-      "loss": 0.8086,
+      "loss": 0.8238,
       "step": 18
     },
     {
       "epoch": 0.01116106030072857,
-      "eval_loss": 0.9415214657783508,
-      "eval_runtime": 47.2291,
-      "eval_samples_per_second": 2.117,
-      "eval_steps_per_second": 2.117,
+      "eval_loss": 0.9594716429710388,
+      "eval_runtime": 47.8775,
+      "eval_samples_per_second": 2.089,
+      "eval_steps_per_second": 2.089,
       "step": 18
     },
     {
       "epoch": 0.011781119206324601,
-      "grad_norm": 0.2330344021320343,
+      "grad_norm": 0.22359193861484528,
       "learning_rate": 8.8e-06,
-      "loss": 0.7858,
+      "loss": 0.8032,
       "step": 19
     },
     {
       "epoch": 0.012401178111920633,
-      "grad_norm": 0.175759956240654,
+      "grad_norm": 0.16142494976520538,
       "learning_rate": 8.000000000000001e-06,
-      "loss": 0.6359,
+      "loss": 0.6508,
       "step": 20
     },
     {
       "epoch": 0.013021237017516665,
-      "grad_norm": 0.21558189392089844,
+      "grad_norm": 0.2063056379556656,
       "learning_rate": 7.2000000000000005e-06,
-      "loss": 0.8966,
+      "loss": 0.9097,
       "step": 21
     },
     {
       "epoch": 0.013021237017516665,
-      "eval_loss": 0.9336352348327637,
-      "eval_runtime": 47.0196,
-      "eval_samples_per_second": 2.127,
-      "eval_steps_per_second": 2.127,
+      "eval_loss": 0.9518673419952393,
+      "eval_runtime": 48.16,
+      "eval_samples_per_second": 2.076,
+      "eval_steps_per_second": 2.076,
       "step": 21
     },
     {
       "epoch": 0.013641295923112695,
-      "grad_norm": 0.2076757252216339,
+      "grad_norm": 0.1905093491077423,
       "learning_rate": 6.4000000000000006e-06,
-      "loss": 0.8719,
+      "loss": 0.8873,
       "step": 22
     },
     {
       "epoch": 0.014261354828708727,
-      "grad_norm": 0.3216860890388489,
+      "grad_norm": 0.28930607438087463,
       "learning_rate": 5.600000000000001e-06,
-      "loss": 1.0026,
+      "loss": 1.0172,
       "step": 23
     },
     {
       "epoch": 0.014881413734304759,
-      "grad_norm": 0.2905324697494507,
+      "grad_norm": 0.2751776874065399,
       "learning_rate": 4.800000000000001e-06,
-      "loss": 0.7967,
+      "loss": 0.8155,
       "step": 24
     },
     {
       "epoch": 0.014881413734304759,
-      "eval_loss": 0.9275597333908081,
-      "eval_runtime": 47.023,
-      "eval_samples_per_second": 2.127,
-      "eval_steps_per_second": 2.127,
+      "eval_loss": 0.946439266204834,
+      "eval_runtime": 47.9311,
+      "eval_samples_per_second": 2.086,
+      "eval_steps_per_second": 2.086,
       "step": 24
     },
     {
       "epoch": 0.01550147263990079,
-      "grad_norm": 0.2005520761013031,
+      "grad_norm": 0.20159322023391724,
       "learning_rate": 4.000000000000001e-06,
-      "loss": 0.8374,
+      "loss": 0.8539,
       "step": 25
     },
     {
       "epoch": 0.016121531545496823,
-      "grad_norm": 0.1872091293334961,
+      "grad_norm": 0.4606916010379791,
       "learning_rate": 3.2000000000000003e-06,
-      "loss": 0.7055,
+      "loss": 0.7225,
       "step": 26
     },
     {
       "epoch": 0.016741590451092855,
-      "grad_norm": 0.1885146200656891,
+      "grad_norm": 0.18102441728115082,
       "learning_rate": 2.4000000000000003e-06,
-      "loss": 0.7527,
+      "loss": 0.7702,
       "step": 27
     },
     {
       "epoch": 0.016741590451092855,
-      "eval_loss": 0.9237930774688721,
-      "eval_runtime": 47.0424,
-      "eval_samples_per_second": 2.126,
-      "eval_steps_per_second": 2.126,
+      "eval_loss": 0.9430135488510132,
+      "eval_runtime": 47.8428,
+      "eval_samples_per_second": 2.09,
+      "eval_steps_per_second": 2.09,
       "step": 27
     },
     {
       "epoch": 0.017361649356688887,
-      "grad_norm": 0.20483697950839996,
+      "grad_norm": 0.2126253843307495,
       "learning_rate": 1.6000000000000001e-06,
-      "loss": 0.6491,
+      "loss": 0.6669,
       "step": 28
     },
     {
       "epoch": 0.01798170826228492,
-      "grad_norm": 0.23465946316719055,
+      "grad_norm": 0.24684152007102966,
       "learning_rate": 8.000000000000001e-07,
-      "loss": 0.9221,
+      "loss": 0.9388,
       "step": 29
     },
     {
       "epoch": 0.01860176716788095,
-      "grad_norm": 0.2271672636270523,
+      "grad_norm": 0.2677140235900879,
       "learning_rate": 0.0,
-      "loss": 0.6819,
+      "loss": 0.7054,
       "step": 30
     },
     {
       "epoch": 0.01860176716788095,
-      "eval_loss": 0.9222295880317688,
-      "eval_runtime": 47.4088,
-      "eval_samples_per_second": 2.109,
-      "eval_steps_per_second": 2.109,
+      "eval_loss": 0.9416863918304443,
+      "eval_runtime": 47.8668,
+      "eval_samples_per_second": 2.089,
+      "eval_steps_per_second": 2.089,
       "step": 30
     }
   ],
@@ -316,7 +316,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 6.35700342245376e+16,
+  "total_flos": 6.450802528167936e+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 sha256:31008ae8f5fa3525c3b5855165b7a47a0cdb47e438b843022563680f47a27521
+oid sha256:b7726c9626d0a377a638b6674780a9e8d32aed5a1fc017243f1823b069fc15ac
 size 5688