Training in progress, step 400, checkpoint

checkpoint-400/adapter_config.json

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

checkpoint-400/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:38710dd67456ffa5ad1c8f42433d3ad70e0b80c2dcabdd6fb489e2b1a85e7936
+oid sha256:d5adfa2b339b645b9226d25986d21f981f96bebdee6f37ac11dfa27026c28829
 size 161533192

checkpoint-400/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:30c60196972226e2f07c5e358fd9d055825a75cb146573cbb8404e600d7cfe73
+oid sha256:9582adeea1a41a07bc22ba561e590ac07067ad872a9776d2d121848fc648237f
 size 323290986

checkpoint-400/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1bad818dce0b02c37af65fb6c646f3c6c96c10be545997b54ce566524dfa2bcc
+oid sha256:bb0935b1d94524502742e7b6f7ccd44a98bf998504cde77342ed19ffb4c6492c
 size 14244

checkpoint-400/trainer_state.json

@@ -10,346 +10,346 @@
   "log_history": [
     {
       "epoch": 0.010604453870625663,
-      "grad_norm": 2.276311159133911,
+      "grad_norm": 2.4981472492218018,
       "learning_rate": 4.982325910215624e-05,
-      "loss": 1.9038,
+      "loss": 1.6893,
       "step": 10
     },
     {
       "epoch": 0.021208907741251327,
-      "grad_norm": 0.7678110003471375,
+      "grad_norm": 0.6994958519935608,
       "learning_rate": 4.964651820431248e-05,
-      "loss": 0.997,
+      "loss": 0.8541,
       "step": 20
     },
     {
       "epoch": 0.031813361611876985,
-      "grad_norm": 0.6726565361022949,
+      "grad_norm": 0.7477335333824158,
       "learning_rate": 4.946977730646872e-05,
-      "loss": 0.8021,
+      "loss": 0.7168,
       "step": 30
     },
     {
       "epoch": 0.042417815482502653,
-      "grad_norm": 0.7318501472473145,
+      "grad_norm": 0.6651986241340637,
       "learning_rate": 4.929303640862496e-05,
-      "loss": 0.7122,
+      "loss": 0.6438,
       "step": 40
     },
     {
       "epoch": 0.053022269353128315,
-      "grad_norm": 0.6708294749259949,
+      "grad_norm": 0.683933436870575,
       "learning_rate": 4.91162955107812e-05,
-      "loss": 0.6783,
+      "loss": 0.6729,
       "step": 50
     },
     {
       "epoch": 0.053022269353128315,
-      "eval_loss": 0.8075584173202515,
-      "eval_runtime": 11.7244,
-      "eval_samples_per_second": 40.94,
-      "eval_steps_per_second": 2.559,
+      "eval_loss": 0.7466596961021423,
+      "eval_runtime": 12.3447,
+      "eval_samples_per_second": 38.883,
+      "eval_steps_per_second": 2.43,
       "step": 50
     },
     {
       "epoch": 0.06362672322375397,
-      "grad_norm": 0.7024317979812622,
+      "grad_norm": 0.6808643341064453,
       "learning_rate": 4.893955461293744e-05,
-      "loss": 0.6244,
+      "loss": 0.6054,
       "step": 60
     },
     {
       "epoch": 0.07423117709437964,
-      "grad_norm": 0.6959078907966614,
+      "grad_norm": 0.6630373001098633,
       "learning_rate": 4.876281371509368e-05,
-      "loss": 0.6117,
+      "loss": 0.582,
       "step": 70
     },
     {
       "epoch": 0.08483563096500531,
-      "grad_norm": 0.7824214696884155,
+      "grad_norm": 0.5939441323280334,
       "learning_rate": 4.858607281724991e-05,
-      "loss": 0.5989,
+      "loss": 0.564,
       "step": 80
     },
     {
       "epoch": 0.09544008483563096,
-      "grad_norm": 0.7792642116546631,
+      "grad_norm": 0.756606936454773,
       "learning_rate": 4.8409331919406155e-05,
-      "loss": 0.6033,
+      "loss": 0.573,
       "step": 90
     },
     {
       "epoch": 0.10604453870625663,
-      "grad_norm": 0.7486705780029297,
+      "grad_norm": 0.7286397218704224,
       "learning_rate": 4.823259102156239e-05,
-      "loss": 0.5904,
+      "loss": 0.5886,
       "step": 100
     },
     {
       "epoch": 0.10604453870625663,
-      "eval_loss": 0.7230511903762817,
-      "eval_runtime": 11.7589,
-      "eval_samples_per_second": 40.82,
-      "eval_steps_per_second": 2.551,
+      "eval_loss": 0.6717663407325745,
+      "eval_runtime": 12.4167,
+      "eval_samples_per_second": 38.658,
+      "eval_steps_per_second": 2.416,
       "step": 100
     },
     {
       "epoch": 0.11664899257688228,
-      "grad_norm": 0.8214134573936462,
+      "grad_norm": 0.7108235359191895,
       "learning_rate": 4.8055850123718633e-05,
-      "loss": 0.5778,
+      "loss": 0.5759,
       "step": 110
     },
     {
       "epoch": 0.12725344644750794,
-      "grad_norm": 0.7937195897102356,
+      "grad_norm": 0.7061440944671631,
       "learning_rate": 4.7879109225874866e-05,
-      "loss": 0.5748,
+      "loss": 0.539,
       "step": 120
     },
     {
       "epoch": 0.1378579003181336,
-      "grad_norm": 0.7781085968017578,
+      "grad_norm": 0.8537445664405823,
       "learning_rate": 4.770236832803111e-05,
-      "loss": 0.5469,
+      "loss": 0.5294,
       "step": 130
     },
     {
       "epoch": 0.14846235418875928,
-      "grad_norm": 0.695203423500061,
+      "grad_norm": 0.740597128868103,
       "learning_rate": 4.7525627430187344e-05,
-      "loss": 0.5511,
+      "loss": 0.5384,
       "step": 140
     },
     {
       "epoch": 0.15906680805938495,
-      "grad_norm": 0.7824985384941101,
+      "grad_norm": 0.6992364525794983,
       "learning_rate": 4.734888653234359e-05,
-      "loss": 0.5258,
+      "loss": 0.5255,
       "step": 150
     },
     {
       "epoch": 0.15906680805938495,
-      "eval_loss": 0.6867014765739441,
-      "eval_runtime": 11.7852,
-      "eval_samples_per_second": 40.729,
-      "eval_steps_per_second": 2.546,
+      "eval_loss": 0.6394088864326477,
+      "eval_runtime": 12.37,
+      "eval_samples_per_second": 38.803,
+      "eval_steps_per_second": 2.425,
       "step": 150
     },
     {
       "epoch": 0.16967126193001061,
-      "grad_norm": 0.7392640113830566,
+      "grad_norm": 0.7426859736442566,
       "learning_rate": 4.717214563449982e-05,
-      "loss": 0.5714,
+      "loss": 0.5483,
       "step": 160
     },
     {
       "epoch": 0.18027571580063625,
-      "grad_norm": 0.8316007852554321,
+      "grad_norm": 0.7601574063301086,
       "learning_rate": 4.699540473665606e-05,
-      "loss": 0.4982,
+      "loss": 0.4993,
       "step": 170
     },
     {
       "epoch": 0.19088016967126192,
-      "grad_norm": 0.8199877738952637,
+      "grad_norm": 0.7388486862182617,
       "learning_rate": 4.681866383881231e-05,
-      "loss": 0.5092,
+      "loss": 0.5175,
       "step": 180
     },
     {
       "epoch": 0.2014846235418876,
-      "grad_norm": 0.7089737057685852,
+      "grad_norm": 0.7000683546066284,
       "learning_rate": 4.664192294096854e-05,
-      "loss": 0.4953,
+      "loss": 0.4936,
       "step": 190
     },
     {
       "epoch": 0.21208907741251326,
-      "grad_norm": 0.7906930446624756,
+      "grad_norm": 0.7684634923934937,
       "learning_rate": 4.6465182043124786e-05,
-      "loss": 0.4974,
+      "loss": 0.5049,
       "step": 200
     },
     {
       "epoch": 0.21208907741251326,
-      "eval_loss": 0.669015645980835,
-      "eval_runtime": 11.8934,
-      "eval_samples_per_second": 40.358,
-      "eval_steps_per_second": 2.522,
+      "eval_loss": 0.6269157528877258,
+      "eval_runtime": 12.3596,
+      "eval_samples_per_second": 38.836,
+      "eval_steps_per_second": 2.427,
       "step": 200
     },
     {
       "epoch": 0.22269353128313893,
-      "grad_norm": 0.9085277318954468,
+      "grad_norm": 0.7945635318756104,
       "learning_rate": 4.628844114528102e-05,
-      "loss": 0.4991,
+      "loss": 0.4957,
       "step": 210
     },
     {
       "epoch": 0.23329798515376457,
-      "grad_norm": 0.7775648832321167,
+      "grad_norm": 0.6900196671485901,
       "learning_rate": 4.6111700247437265e-05,
-      "loss": 0.5024,
+      "loss": 0.5009,
       "step": 220
     },
     {
       "epoch": 0.24390243902439024,
-      "grad_norm": 0.8324123620986938,
+      "grad_norm": 0.7506561875343323,
       "learning_rate": 4.59349593495935e-05,
-      "loss": 0.5187,
+      "loss": 0.509,
       "step": 230
     },
     {
       "epoch": 0.2545068928950159,
-      "grad_norm": 0.8002191781997681,
+      "grad_norm": 0.7476117014884949,
       "learning_rate": 4.5758218451749736e-05,
-      "loss": 0.4785,
+      "loss": 0.4755,
       "step": 240
     },
     {
       "epoch": 0.2651113467656416,
-      "grad_norm": 0.8004704117774963,
+      "grad_norm": 0.7749964594841003,
       "learning_rate": 4.5581477553905975e-05,
-      "loss": 0.5073,
+      "loss": 0.5163,
       "step": 250
     },
     {
       "epoch": 0.2651113467656416,
-      "eval_loss": 0.6479213833808899,
-      "eval_runtime": 11.882,
-      "eval_samples_per_second": 40.397,
-      "eval_steps_per_second": 2.525,
+      "eval_loss": 0.6084603667259216,
+      "eval_runtime": 12.3487,
+      "eval_samples_per_second": 38.87,
+      "eval_steps_per_second": 2.429,
       "step": 250
     },
     {
       "epoch": 0.2757158006362672,
-      "grad_norm": 0.893301248550415,
+      "grad_norm": 0.7424341440200806,
       "learning_rate": 4.5404736656062215e-05,
-      "loss": 0.477,
+      "loss": 0.4766,
       "step": 260
     },
     {
       "epoch": 0.2863202545068929,
-      "grad_norm": 0.8550834059715271,
+      "grad_norm": 0.7572893500328064,
       "learning_rate": 4.5227995758218454e-05,
-      "loss": 0.4923,
+      "loss": 0.4909,
       "step": 270
     },
     {
       "epoch": 0.29692470837751855,
-      "grad_norm": 0.8929152488708496,
+      "grad_norm": 0.7275112271308899,
       "learning_rate": 4.505125486037469e-05,
-      "loss": 0.4624,
+      "loss": 0.486,
       "step": 280
     },
     {
       "epoch": 0.3075291622481442,
-      "grad_norm": 0.9584974050521851,
+      "grad_norm": 0.7402988076210022,
       "learning_rate": 4.487451396253093e-05,
-      "loss": 0.4801,
+      "loss": 0.4891,
       "step": 290
     },
     {
       "epoch": 0.3181336161187699,
-      "grad_norm": 0.7924031615257263,
+      "grad_norm": 0.7442552447319031,
       "learning_rate": 4.469777306468717e-05,
-      "loss": 0.4942,
+      "loss": 0.4794,
       "step": 300
     },
     {
       "epoch": 0.3181336161187699,
-      "eval_loss": 0.6371316313743591,
-      "eval_runtime": 11.8575,
-      "eval_samples_per_second": 40.481,
-      "eval_steps_per_second": 2.53,
+      "eval_loss": 0.5998384952545166,
+      "eval_runtime": 12.3533,
+      "eval_samples_per_second": 38.856,
+      "eval_steps_per_second": 2.428,
       "step": 300
     },
     {
       "epoch": 0.32873806998939553,
-      "grad_norm": 0.9247742295265198,
+      "grad_norm": 0.7272119522094727,
       "learning_rate": 4.452103216684341e-05,
-      "loss": 0.5234,
+      "loss": 0.5141,
       "step": 310
     },
     {
       "epoch": 0.33934252386002123,
-      "grad_norm": 0.8617805242538452,
+      "grad_norm": 0.8105739951133728,
       "learning_rate": 4.434429126899965e-05,
-      "loss": 0.4808,
+      "loss": 0.4897,
       "step": 320
     },
     {
       "epoch": 0.34994697773064687,
-      "grad_norm": 0.8334285616874695,
+      "grad_norm": 0.7440537810325623,
       "learning_rate": 4.416755037115589e-05,
-      "loss": 0.4758,
+      "loss": 0.4721,
       "step": 330
     },
     {
       "epoch": 0.3605514316012725,
-      "grad_norm": 0.9079923629760742,
+      "grad_norm": 0.7973731756210327,
       "learning_rate": 4.399080947331213e-05,
-      "loss": 0.463,
+      "loss": 0.4643,
       "step": 340
     },
     {
       "epoch": 0.3711558854718982,
-      "grad_norm": 0.7632007002830505,
+      "grad_norm": 0.7304461598396301,
       "learning_rate": 4.381406857546837e-05,
-      "loss": 0.4828,
+      "loss": 0.4907,
       "step": 350
     },
     {
       "epoch": 0.3711558854718982,
-      "eval_loss": 0.6212183237075806,
-      "eval_runtime": 11.8497,
-      "eval_samples_per_second": 40.507,
-      "eval_steps_per_second": 2.532,
+      "eval_loss": 0.5852673649787903,
+      "eval_runtime": 12.3481,
+      "eval_samples_per_second": 38.872,
+      "eval_steps_per_second": 2.43,
       "step": 350
     },
     {
       "epoch": 0.38176033934252385,
-      "grad_norm": 0.8430343270301819,
+      "grad_norm": 0.68657386302948,
       "learning_rate": 4.3637327677624606e-05,
-      "loss": 0.4711,
+      "loss": 0.4734,
       "step": 360
     },
     {
       "epoch": 0.39236479321314954,
-      "grad_norm": 0.9495126605033875,
+      "grad_norm": 0.8211691379547119,
       "learning_rate": 4.3460586779780846e-05,
-      "loss": 0.4528,
+      "loss": 0.457,
       "step": 370
     },
     {
       "epoch": 0.4029692470837752,
-      "grad_norm": 0.8903242945671082,
+      "grad_norm": 0.7916038036346436,
       "learning_rate": 4.328384588193708e-05,
-      "loss": 0.468,
+      "loss": 0.4649,
       "step": 380
     },
     {
       "epoch": 0.4135737009544008,
-      "grad_norm": 0.9306944608688354,
+      "grad_norm": 0.7529716491699219,
       "learning_rate": 4.3107104984093324e-05,
-      "loss": 0.4299,
+      "loss": 0.4437,
       "step": 390
     },
     {
       "epoch": 0.4241781548250265,
-      "grad_norm": 0.7949309349060059,
+      "grad_norm": 0.7294536828994751,
       "learning_rate": 4.2930364086249556e-05,
-      "loss": 0.4589,
+      "loss": 0.4668,
       "step": 400
     },
     {
       "epoch": 0.4241781548250265,
-      "eval_loss": 0.6148895621299744,
-      "eval_runtime": 11.8536,
-      "eval_samples_per_second": 40.494,
-      "eval_steps_per_second": 2.531,
+      "eval_loss": 0.5796846747398376,
+      "eval_runtime": 12.3341,
+      "eval_samples_per_second": 38.916,
+      "eval_steps_per_second": 2.432,
       "step": 400
     }
   ],
@@ -370,7 +370,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 4.142486912935526e+16,
+  "total_flos": 4.113610607537357e+16,
   "train_batch_size": 16,
   "trial_name": null,
   "trial_params": null

checkpoint-400/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:53a8c021c0523ab617f7d132f6e1d21f2075e97fdb7689e840960f38ef5cb15e
+oid sha256:c5186a048b8d2af8a229717dd049aeeb2812c6aef0d473c9569aaea2283f4a48
 size 5880