Training in progress, step 100, checkpoint

checkpoint-100/adapter_config.json

@@ -24,12 +24,12 @@
   "revision": null,
   "target_modules": [
     "gate_proj",
-    "q_proj",
+    "v_proj",
     "k_proj",
-    "up_proj",
+    "q_proj",
     "o_proj",
-    "down_proj",
-    "v_proj"
+    "up_proj",
+    "down_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

checkpoint-100/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c7a8bc21fa2e8a47b90a2816d7be2e2bf2dbb623a1119ece2f43a082c62d469a
+oid sha256:c81f7e0e47d5acc29d6efedf7aabffa931304887d58bb084174adf8e936b6b43
 size 216151256

checkpoint-100/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:10ea843b4f8a0fac716dca25b8e9c16d7bb8ff022d4bb6695c5aacaa7518c72e
+oid sha256:4c818751d33036cf8f4d360814fe89ffbc54a01f0c8f4f12f93ad68ae9353099
 size 432640054

checkpoint-100/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bc0f4b4def19d80238543e9204366e45f9cfc570f06744d235af389f5feb7e13
+oid sha256:40a12c8bf4a99e5b8cf8700f3f0607f02765e51239fdbfb9883f71b9b2047d61
 size 14244

checkpoint-100/trainer_state.json

@@ -2,7 +2,7 @@
   "best_metric": null,
   "best_model_checkpoint": null,
   "epoch": 0.12658227848101267,
-  "eval_steps": 50,
+  "eval_steps": 10,
   "global_step": 100,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
@@ -10,88 +10,152 @@
   "log_history": [
     {
       "epoch": 0.012658227848101266,
-      "grad_norm": 16.027029037475586,
+      "grad_norm": 12.934767723083496,
       "learning_rate": 4.9789029535864986e-05,
-      "loss": 2.6925,
+      "loss": 2.6869,
+      "step": 10
+    },
+    {
+      "epoch": 0.012658227848101266,
+      "eval_loss": 1.651185393333435,
+      "eval_runtime": 11.831,
+      "eval_samples_per_second": 40.571,
+      "eval_steps_per_second": 2.536,
       "step": 10
     },
     {
       "epoch": 0.02531645569620253,
-      "grad_norm": 1.0948777198791504,
+      "grad_norm": 1.104798674583435,
       "learning_rate": 4.957805907172996e-05,
-      "loss": 1.386,
+      "loss": 1.3694,
+      "step": 20
+    },
+    {
+      "epoch": 0.02531645569620253,
+      "eval_loss": 1.2200205326080322,
+      "eval_runtime": 11.8928,
+      "eval_samples_per_second": 40.361,
+      "eval_steps_per_second": 2.523,
       "step": 20
     },
     {
       "epoch": 0.0379746835443038,
-      "grad_norm": 1.1341983079910278,
+      "grad_norm": 1.1069142818450928,
       "learning_rate": 4.936708860759494e-05,
-      "loss": 1.108,
+      "loss": 1.1029,
+      "step": 30
+    },
+    {
+      "epoch": 0.0379746835443038,
+      "eval_loss": 1.0691736936569214,
+      "eval_runtime": 11.9127,
+      "eval_samples_per_second": 40.293,
+      "eval_steps_per_second": 2.518,
       "step": 30
     },
     {
       "epoch": 0.05063291139240506,
-      "grad_norm": 1.152009129524231,
+      "grad_norm": 1.1594161987304688,
       "learning_rate": 4.9156118143459915e-05,
-      "loss": 0.9436,
+      "loss": 0.9395,
+      "step": 40
+    },
+    {
+      "epoch": 0.05063291139240506,
+      "eval_loss": 0.9162012934684753,
+      "eval_runtime": 11.9373,
+      "eval_samples_per_second": 40.21,
+      "eval_steps_per_second": 2.513,
       "step": 40
     },
     {
       "epoch": 0.06329113924050633,
-      "grad_norm": 1.0990614891052246,
+      "grad_norm": 1.1133538484573364,
       "learning_rate": 4.89451476793249e-05,
-      "loss": 0.7499,
+      "loss": 0.7489,
       "step": 50
     },
     {
       "epoch": 0.06329113924050633,
-      "eval_loss": 0.7157873511314392,
-      "eval_runtime": 12.1594,
-      "eval_samples_per_second": 39.476,
-      "eval_steps_per_second": 2.467,
+      "eval_loss": 0.713701605796814,
+      "eval_runtime": 11.9661,
+      "eval_samples_per_second": 40.113,
+      "eval_steps_per_second": 2.507,
       "step": 50
     },
     {
       "epoch": 0.0759493670886076,
-      "grad_norm": 1.0284477472305298,
+      "grad_norm": 1.0406183004379272,
       "learning_rate": 4.8734177215189874e-05,
-      "loss": 0.6095,
+      "loss": 0.6096,
+      "step": 60
+    },
+    {
+      "epoch": 0.0759493670886076,
+      "eval_loss": 0.6309535503387451,
+      "eval_runtime": 11.9895,
+      "eval_samples_per_second": 40.035,
+      "eval_steps_per_second": 2.502,
       "step": 60
     },
     {
       "epoch": 0.08860759493670886,
-      "grad_norm": 0.8609589338302612,
+      "grad_norm": 0.8599340915679932,
       "learning_rate": 4.852320675105486e-05,
-      "loss": 0.5355,
+      "loss": 0.5357,
+      "step": 70
+    },
+    {
+      "epoch": 0.08860759493670886,
+      "eval_loss": 0.6159886717796326,
+      "eval_runtime": 12.0107,
+      "eval_samples_per_second": 39.965,
+      "eval_steps_per_second": 2.498,
       "step": 70
     },
     {
       "epoch": 0.10126582278481013,
-      "grad_norm": 0.9162376523017883,
+      "grad_norm": 0.9128267168998718,
       "learning_rate": 4.8312236286919834e-05,
-      "loss": 0.5705,
+      "loss": 0.5703,
+      "step": 80
+    },
+    {
+      "epoch": 0.10126582278481013,
+      "eval_loss": 0.5933937430381775,
+      "eval_runtime": 11.9716,
+      "eval_samples_per_second": 40.095,
+      "eval_steps_per_second": 2.506,
       "step": 80
     },
     {
       "epoch": 0.11392405063291139,
-      "grad_norm": 0.9415847659111023,
+      "grad_norm": 0.9396541118621826,
       "learning_rate": 4.810126582278481e-05,
-      "loss": 0.5449,
+      "loss": 0.5445,
+      "step": 90
+    },
+    {
+      "epoch": 0.11392405063291139,
+      "eval_loss": 0.5727818608283997,
+      "eval_runtime": 11.9685,
+      "eval_samples_per_second": 40.105,
+      "eval_steps_per_second": 2.507,
       "step": 90
     },
     {
       "epoch": 0.12658227848101267,
-      "grad_norm": 0.8756884336471558,
+      "grad_norm": 0.8805290460586548,
       "learning_rate": 4.789029535864979e-05,
-      "loss": 0.5157,
+      "loss": 0.5151,
       "step": 100
     },
     {
       "epoch": 0.12658227848101267,
-      "eval_loss": 0.563517153263092,
-      "eval_runtime": 12.1236,
-      "eval_samples_per_second": 39.592,
-      "eval_steps_per_second": 2.475,
+      "eval_loss": 0.5640087127685547,
+      "eval_runtime": 11.9824,
+      "eval_samples_per_second": 40.059,
+      "eval_steps_per_second": 2.504,
       "step": 100
     }
   ],
@@ -99,7 +163,7 @@
   "max_steps": 2370,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 3,
-  "save_steps": 100,
+  "save_steps": 10,
   "stateful_callbacks": {
     "TrainerControl": {
       "args": {

checkpoint-100/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b8ef5fa4aad3a350c14df025074931ad8a003d4b851f4886f3b2f66ae6653e4b
+oid sha256:f1e7aae8b855413d55586dd498c7d7d805796f0c02067ce9d8ccb1ef37f72d29
 size 5880