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 sha256:7696ce0eabde7eb93d1bd677680a947fd0923b0334870c81bca1f2af8380ff2e
+oid sha256:50a02227c72203365ffc1ccba8c36d11224669ad85c6c6c25ae975396d5a9de5
 size 83945296

checkpoint-30/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:94ea4d145246659349040ad95bd6972a9e718f69bd087672f8328627f4166cd3
+oid sha256:306571fa180a4fc75435489b12fd03ef84a6d2332d52e5df827bfec1029e201c
 size 168149074

checkpoint-30/trainer_state.json

@@ -10,292 +10,292 @@
   "log_history": [
     {
       "epoch": 0.0006200589055960316,
-      "grad_norm": 0.7731499075889587,
+      "grad_norm": 0.7607094645500183,
       "learning_rate": 4.000000000000001e-06,
-      "loss": 1.166,
+      "loss": 1.1893,
       "step": 1
     },
     {
       "epoch": 0.0012401178111920632,
-      "grad_norm": 0.346652090549469,
+      "grad_norm": 0.3518824279308319,
       "learning_rate": 8.000000000000001e-06,
-      "loss": 0.9594,
+      "loss": 0.9752,
       "step": 2
     },
     {
       "epoch": 0.0018601767167880949,
-      "grad_norm": 0.39937612414360046,
+      "grad_norm": 0.4024870693683624,
       "learning_rate": 1.2e-05,
-      "loss": 0.828,
+      "loss": 0.8467,
       "step": 3
     },
     {
       "epoch": 0.0018601767167880949,
-      "eval_loss": 1.0491173267364502,
-      "eval_runtime": 26.0459,
-      "eval_samples_per_second": 3.839,
-      "eval_steps_per_second": 3.839,
+      "eval_loss": 1.0757174491882324,
+      "eval_runtime": 26.414,
+      "eval_samples_per_second": 3.786,
+      "eval_steps_per_second": 3.786,
       "step": 3
     },
     {
       "epoch": 0.0024802356223841263,
-      "grad_norm": 0.3457990288734436,
+      "grad_norm": 0.36653462052345276,
       "learning_rate": 1.6000000000000003e-05,
-      "loss": 0.988,
+      "loss": 1.0056,
       "step": 4
     },
     {
       "epoch": 0.0031002945279801583,
-      "grad_norm": 0.513918399810791,
+      "grad_norm": 0.47098514437675476,
       "learning_rate": 2e-05,
-      "loss": 0.8319,
+      "loss": 0.8545,
       "step": 5
     },
     {
       "epoch": 0.0037203534335761897,
-      "grad_norm": 0.47683727741241455,
+      "grad_norm": 0.5030912160873413,
       "learning_rate": 1.9200000000000003e-05,
-      "loss": 0.9366,
+      "loss": 0.9553,
       "step": 6
     },
     {
       "epoch": 0.0037203534335761897,
-      "eval_loss": 1.0431987047195435,
-      "eval_runtime": 26.2352,
-      "eval_samples_per_second": 3.812,
-      "eval_steps_per_second": 3.812,
+      "eval_loss": 1.0662517547607422,
+      "eval_runtime": 26.7124,
+      "eval_samples_per_second": 3.744,
+      "eval_steps_per_second": 3.744,
       "step": 6
     },
     {
       "epoch": 0.004340412339172222,
-      "grad_norm": 0.4349649250507355,
+      "grad_norm": 0.4823940694332123,
       "learning_rate": 1.8400000000000003e-05,
-      "loss": 0.9076,
+      "loss": 0.9266,
       "step": 7
     },
     {
       "epoch": 0.004960471244768253,
-      "grad_norm": 0.5443339943885803,
+      "grad_norm": 0.5647093653678894,
       "learning_rate": 1.76e-05,
-      "loss": 0.9829,
+      "loss": 1.0013,
       "step": 8
     },
     {
       "epoch": 0.005580530150364285,
-      "grad_norm": 0.3990451395511627,
+      "grad_norm": 0.4058099687099457,
       "learning_rate": 1.6800000000000002e-05,
-      "loss": 0.9045,
+      "loss": 0.9206,
       "step": 9
     },
     {
       "epoch": 0.005580530150364285,
-      "eval_loss": 1.0292819738388062,
-      "eval_runtime": 26.3477,
-      "eval_samples_per_second": 3.795,
-      "eval_steps_per_second": 3.795,
+      "eval_loss": 1.0476710796356201,
+      "eval_runtime": 26.7708,
+      "eval_samples_per_second": 3.735,
+      "eval_steps_per_second": 3.735,
       "step": 9
     },
     {
       "epoch": 0.0062005890559603165,
-      "grad_norm": 0.44750896096229553,
+      "grad_norm": 0.44608208537101746,
       "learning_rate": 1.6000000000000003e-05,
-      "loss": 1.0768,
+      "loss": 1.0937,
       "step": 10
     },
     {
       "epoch": 0.0068206479615563476,
-      "grad_norm": 0.6865904331207275,
+      "grad_norm": 0.7088199853897095,
       "learning_rate": 1.5200000000000002e-05,
-      "loss": 0.8717,
+      "loss": 0.8915,
       "step": 11
     },
     {
       "epoch": 0.0074407068671523795,
-      "grad_norm": 0.548491895198822,
+      "grad_norm": 0.5858705639839172,
       "learning_rate": 1.4400000000000001e-05,
-      "loss": 1.1758,
+      "loss": 1.186,
       "step": 12
     },
     {
       "epoch": 0.0074407068671523795,
-      "eval_loss": 1.0153621435165405,
-      "eval_runtime": 26.5025,
-      "eval_samples_per_second": 3.773,
-      "eval_steps_per_second": 3.773,
+      "eval_loss": 1.0301631689071655,
+      "eval_runtime": 27.1639,
+      "eval_samples_per_second": 3.681,
+      "eval_steps_per_second": 3.681,
       "step": 12
     },
     {
       "epoch": 0.008060765772748411,
-      "grad_norm": 0.5520397424697876,
+      "grad_norm": 0.5912883281707764,
       "learning_rate": 1.3600000000000002e-05,
-      "loss": 0.992,
+      "loss": 1.0082,
       "step": 13
     },
     {
       "epoch": 0.008680824678344443,
-      "grad_norm": 0.381074458360672,
+      "grad_norm": 0.42219021916389465,
       "learning_rate": 1.2800000000000001e-05,
-      "loss": 0.8504,
+      "loss": 0.865,
       "step": 14
     },
     {
       "epoch": 0.009300883583940475,
-      "grad_norm": 0.41114744544029236,
+      "grad_norm": 0.4253869652748108,
       "learning_rate": 1.2e-05,
-      "loss": 0.8685,
+      "loss": 0.8849,
       "step": 15
     },
     {
       "epoch": 0.009300883583940475,
-      "eval_loss": 1.0006165504455566,
-      "eval_runtime": 26.4086,
-      "eval_samples_per_second": 3.787,
-      "eval_steps_per_second": 3.787,
+      "eval_loss": 1.012740135192871,
+      "eval_runtime": 27.032,
+      "eval_samples_per_second": 3.699,
+      "eval_steps_per_second": 3.699,
       "step": 15
     },
     {
       "epoch": 0.009920942489536505,
-      "grad_norm": 0.5197786688804626,
+      "grad_norm": 0.5621253252029419,
       "learning_rate": 1.1200000000000001e-05,
-      "loss": 0.8176,
+      "loss": 0.8289,
       "step": 16
     },
     {
       "epoch": 0.010541001395132537,
-      "grad_norm": 0.3905779719352722,
+      "grad_norm": 0.37459954619407654,
       "learning_rate": 1.04e-05,
-      "loss": 0.9055,
+      "loss": 0.9139,
       "step": 17
     },
     {
       "epoch": 0.01116106030072857,
-      "grad_norm": 0.43444159626960754,
+      "grad_norm": 0.448284387588501,
       "learning_rate": 9.600000000000001e-06,
-      "loss": 0.8452,
+      "loss": 0.8571,
       "step": 18
     },
     {
       "epoch": 0.01116106030072857,
-      "eval_loss": 0.9866559505462646,
-      "eval_runtime": 26.3639,
-      "eval_samples_per_second": 3.793,
-      "eval_steps_per_second": 3.793,
+      "eval_loss": 0.99790358543396,
+      "eval_runtime": 26.9342,
+      "eval_samples_per_second": 3.713,
+      "eval_steps_per_second": 3.713,
       "step": 18
     },
     {
       "epoch": 0.011781119206324601,
-      "grad_norm": 0.5130135416984558,
+      "grad_norm": 0.5347204208374023,
       "learning_rate": 8.8e-06,
-      "loss": 0.8276,
+      "loss": 0.8396,
       "step": 19
     },
     {
       "epoch": 0.012401178111920633,
-      "grad_norm": 0.35828444361686707,
+      "grad_norm": 0.36863428354263306,
       "learning_rate": 8.000000000000001e-06,
-      "loss": 0.6587,
+      "loss": 0.6704,
       "step": 20
     },
     {
       "epoch": 0.013021237017516665,
-      "grad_norm": 0.4369220435619354,
+      "grad_norm": 0.4834197163581848,
       "learning_rate": 7.2000000000000005e-06,
-      "loss": 0.9372,
+      "loss": 0.9445,
       "step": 21
     },
     {
       "epoch": 0.013021237017516665,
-      "eval_loss": 0.9746317267417908,
-      "eval_runtime": 26.2985,
-      "eval_samples_per_second": 3.803,
-      "eval_steps_per_second": 3.803,
+      "eval_loss": 0.9843659996986389,
+      "eval_runtime": 27.0084,
+      "eval_samples_per_second": 3.703,
+      "eval_steps_per_second": 3.703,
       "step": 21
     },
     {
       "epoch": 0.013641295923112695,
-      "grad_norm": 0.4957813024520874,
+      "grad_norm": 0.5544454455375671,
       "learning_rate": 6.4000000000000006e-06,
-      "loss": 0.9153,
+      "loss": 0.9225,
       "step": 22
     },
     {
       "epoch": 0.014261354828708727,
-      "grad_norm": 0.5806369185447693,
+      "grad_norm": 0.6251612901687622,
       "learning_rate": 5.600000000000001e-06,
-      "loss": 1.0175,
+      "loss": 1.0261,
       "step": 23
     },
     {
       "epoch": 0.014881413734304759,
-      "grad_norm": 0.42675939202308655,
+      "grad_norm": 0.5012588500976562,
       "learning_rate": 4.800000000000001e-06,
-      "loss": 0.8421,
+      "loss": 0.8526,
       "step": 24
     },
     {
       "epoch": 0.014881413734304759,
-      "eval_loss": 0.9651661515235901,
-      "eval_runtime": 26.4248,
-      "eval_samples_per_second": 3.784,
-      "eval_steps_per_second": 3.784,
+      "eval_loss": 0.9725684523582458,
+      "eval_runtime": 26.7432,
+      "eval_samples_per_second": 3.739,
+      "eval_steps_per_second": 3.739,
       "step": 24
     },
     {
       "epoch": 0.01550147263990079,
-      "grad_norm": 0.554572343826294,
+      "grad_norm": 0.5470224618911743,
       "learning_rate": 4.000000000000001e-06,
-      "loss": 0.8981,
+      "loss": 0.9036,
       "step": 25
     },
     {
       "epoch": 0.016121531545496823,
-      "grad_norm": 0.4890541136264801,
+      "grad_norm": 0.7129680514335632,
       "learning_rate": 3.2000000000000003e-06,
-      "loss": 0.7742,
+      "loss": 0.7822,
       "step": 26
     },
     {
       "epoch": 0.016741590451092855,
-      "grad_norm": 0.4652097821235657,
+      "grad_norm": 0.6765540838241577,
       "learning_rate": 2.4000000000000003e-06,
-      "loss": 0.7888,
+      "loss": 0.7982,
       "step": 27
     },
     {
       "epoch": 0.016741590451092855,
-      "eval_loss": 0.9591427445411682,
-      "eval_runtime": 26.4791,
-      "eval_samples_per_second": 3.777,
-      "eval_steps_per_second": 3.777,
+      "eval_loss": 0.965038537979126,
+      "eval_runtime": 27.034,
+      "eval_samples_per_second": 3.699,
+      "eval_steps_per_second": 3.699,
       "step": 27
     },
     {
       "epoch": 0.017361649356688887,
-      "grad_norm": 0.3710659444332123,
+      "grad_norm": 0.6408655643463135,
       "learning_rate": 1.6000000000000001e-06,
-      "loss": 0.6674,
+      "loss": 0.6792,
       "step": 28
     },
     {
       "epoch": 0.01798170826228492,
-      "grad_norm": 0.5322249531745911,
+      "grad_norm": 0.7210974097251892,
       "learning_rate": 8.000000000000001e-07,
-      "loss": 0.9623,
+      "loss": 0.9673,
       "step": 29
     },
     {
       "epoch": 0.01860176716788095,
-      "grad_norm": 0.6200012564659119,
+      "grad_norm": 0.6966854929924011,
       "learning_rate": 0.0,
-      "loss": 0.699,
+      "loss": 0.7089,
       "step": 30
     },
     {
       "epoch": 0.01860176716788095,
-      "eval_loss": 0.9562991261482239,
-      "eval_runtime": 26.5068,
-      "eval_samples_per_second": 3.773,
-      "eval_steps_per_second": 3.773,
+      "eval_loss": 0.9616574048995972,
+      "eval_runtime": 27.0027,
+      "eval_samples_per_second": 3.703,
+      "eval_steps_per_second": 3.703,
       "step": 30
     }
   ],
@@ -316,7 +316,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 3.37446133456896e+16,
+  "total_flos": 3.42368069566464e+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:cbe2fb31cf4c04f893e5299faa77267af76b8447f922d82c2d6c0200e1fe225f
+oid sha256:527a35da8257b0136b41661334383dc9a90d4f2271ec7e0b37f6fe024a71f4c9
 size 5688