Training in progress, step 70, checkpoint

checkpoint-70/adapter_config.json

@@ -23,11 +23,11 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "k_proj",
     "v_proj",
-    "o_proj",
-    "gate_proj",
+    "k_proj",
     "q_proj",
+    "gate_proj",
+    "o_proj",
     "down_proj",
     "up_proj"
   ],

checkpoint-70/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:497c1a2809b16acf1e3de74649a6cd989d7bd32c4b1ca6d8c37789503276c86d
+oid sha256:47885d78c898d6ec34947040784edc7617a4de70342cc80bd6a7688601dbd6ed
 size 145287696

checkpoint-70/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c2386bd2314f4c75cce6957016cf2ab951659d522db66de5ec387b34a52be8cf
+oid sha256:09f2bfd7a1ce01304f4307de61e6f2afea404747362f1b8e7da0f4cb7de6323b
 size 290833618

checkpoint-70/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9316543e09853a805290cd6f5a5bb7f3480626115c4cbde98ef584fa48396437
+oid sha256:174965c0191f95cc5c3ff4ffb7dd99e0ad61f9f08edf482de6218a0cdc8be32f
 size 1064

checkpoint-70/trainer_state.json

@@ -10,114 +10,114 @@
   "log_history": [
     {
       "epoch": 0.021739130434782608,
-      "grad_norm": 2.138256549835205,
-      "learning_rate": 4.5e-05,
-      "loss": 2.3657,
+      "grad_norm": 2.2049312591552734,
+      "learning_rate": 4.963768115942029e-05,
+      "loss": 2.3459,
       "step": 10
     },
     {
       "epoch": 0.021739130434782608,
-      "eval_loss": 1.5849242210388184,
-      "eval_runtime": 9.6002,
-      "eval_samples_per_second": 49.999,
-      "eval_steps_per_second": 1.562,
+      "eval_loss": 1.5330805778503418,
+      "eval_runtime": 9.6608,
+      "eval_samples_per_second": 49.685,
+      "eval_steps_per_second": 1.553,
       "step": 10
     },
     {
       "epoch": 0.043478260869565216,
-      "grad_norm": 1.1203280687332153,
-      "learning_rate": 4e-05,
-      "loss": 1.2125,
+      "grad_norm": 1.1927831172943115,
+      "learning_rate": 4.9275362318840584e-05,
+      "loss": 1.1669,
       "step": 20
     },
     {
       "epoch": 0.043478260869565216,
-      "eval_loss": 1.04584538936615,
-      "eval_runtime": 9.677,
-      "eval_samples_per_second": 49.602,
-      "eval_steps_per_second": 1.55,
+      "eval_loss": 1.0271039009094238,
+      "eval_runtime": 9.7215,
+      "eval_samples_per_second": 49.375,
+      "eval_steps_per_second": 1.543,
       "step": 20
     },
     {
       "epoch": 0.06521739130434782,
-      "grad_norm": 1.064527988433838,
-      "learning_rate": 3.5e-05,
-      "loss": 0.867,
+      "grad_norm": 1.0509072542190552,
+      "learning_rate": 4.891304347826087e-05,
+      "loss": 0.8327,
       "step": 30
     },
     {
       "epoch": 0.06521739130434782,
-      "eval_loss": 0.9389117360115051,
-      "eval_runtime": 9.7352,
-      "eval_samples_per_second": 49.305,
-      "eval_steps_per_second": 1.541,
+      "eval_loss": 0.9039635062217712,
+      "eval_runtime": 9.7818,
+      "eval_samples_per_second": 49.071,
+      "eval_steps_per_second": 1.533,
       "step": 30
     },
     {
       "epoch": 0.08695652173913043,
-      "grad_norm": 1.1056835651397705,
-      "learning_rate": 3e-05,
-      "loss": 0.7879,
+      "grad_norm": 0.8860757350921631,
+      "learning_rate": 4.855072463768116e-05,
+      "loss": 0.731,
       "step": 40
     },
     {
       "epoch": 0.08695652173913043,
-      "eval_loss": 0.8724283576011658,
-      "eval_runtime": 9.8116,
-      "eval_samples_per_second": 48.922,
+      "eval_loss": 0.8067704439163208,
+      "eval_runtime": 9.8118,
+      "eval_samples_per_second": 48.921,
       "eval_steps_per_second": 1.529,
       "step": 40
     },
     {
       "epoch": 0.10869565217391304,
-      "grad_norm": 0.8273067474365234,
-      "learning_rate": 2.5e-05,
-      "loss": 0.7057,
+      "grad_norm": 0.57489013671875,
+      "learning_rate": 4.818840579710145e-05,
+      "loss": 0.6501,
       "step": 50
     },
     {
       "epoch": 0.10869565217391304,
-      "eval_loss": 0.8029478788375854,
-      "eval_runtime": 9.8368,
-      "eval_samples_per_second": 48.797,
-      "eval_steps_per_second": 1.525,
+      "eval_loss": 0.7662845849990845,
+      "eval_runtime": 9.8455,
+      "eval_samples_per_second": 48.753,
+      "eval_steps_per_second": 1.524,
       "step": 50
     },
     {
       "epoch": 0.13043478260869565,
-      "grad_norm": 0.5044384598731995,
-      "learning_rate": 2e-05,
-      "loss": 0.6629,
+      "grad_norm": 0.5091506242752075,
+      "learning_rate": 4.782608695652174e-05,
+      "loss": 0.6288,
       "step": 60
     },
     {
       "epoch": 0.13043478260869565,
-      "eval_loss": 0.7792357802391052,
-      "eval_runtime": 9.8462,
-      "eval_samples_per_second": 48.75,
-      "eval_steps_per_second": 1.523,
+      "eval_loss": 0.7400590181350708,
+      "eval_runtime": 9.8714,
+      "eval_samples_per_second": 48.626,
+      "eval_steps_per_second": 1.52,
       "step": 60
     },
     {
       "epoch": 0.15217391304347827,
-      "grad_norm": 0.5917642116546631,
-      "learning_rate": 1.5e-05,
-      "loss": 0.623,
+      "grad_norm": 0.5590857863426208,
+      "learning_rate": 4.746376811594203e-05,
+      "loss": 0.5838,
       "step": 70
     },
     {
       "epoch": 0.15217391304347827,
-      "eval_loss": 0.7707605957984924,
-      "eval_runtime": 9.8356,
-      "eval_samples_per_second": 48.803,
-      "eval_steps_per_second": 1.525,
+      "eval_loss": 0.7244360446929932,
+      "eval_runtime": 9.8787,
+      "eval_samples_per_second": 48.589,
+      "eval_steps_per_second": 1.518,
       "step": 70
     }
   ],
   "logging_steps": 10,
-  "max_steps": 100,
+  "max_steps": 1380,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 1,
+  "num_train_epochs": 3,
   "save_steps": 10,
   "stateful_callbacks": {
     "TrainerControl": {

checkpoint-70/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:17a1a7c75c85fbba54f1e97b2cee73c167834890dd0df62950989daaf8975814
+oid sha256:1cdf2cbd96a7d8012e1f3da0569783099ba26a3ecbec680ff36cde09793d0889
 size 5816