medquad

adapter_config.json

@@ -22,12 +22,12 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "down_proj",
-    "up_proj",
     "k_proj",
-    "o_proj",
     "v_proj",
+    "up_proj",
     "q_proj",
+    "down_proj",
+    "o_proj",
     "gate_proj"
   ],
   "task_type": "CAUSAL_LM",

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d680bb0c04772bf8f307ff155fdada2c78e6740d40f29e832d239e9f0eca7ebf
+oid sha256:254effd1887e12c69235899e5dffc7f062a2bca08827378ed22e910f52fd679e
 size 19644664

all_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 5.0,
     "total_flos": 549306960445440.0,
-    "train_loss": 2.4234995365142824,
-    "train_runtime": 54.6732,
+    "train_loss": 2.424940276145935,
+    "train_runtime": 55.452,
     "train_samples": 8643,
-    "train_samples_per_second": 1.463,
-    "train_steps_per_second": 0.183
+    "train_samples_per_second": 1.443,
+    "train_steps_per_second": 0.18
 }

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 5.0,
     "total_flos": 549306960445440.0,
-    "train_loss": 2.4234995365142824,
-    "train_runtime": 54.6732,
+    "train_loss": 2.424940276145935,
+    "train_runtime": 55.452,
     "train_samples": 8643,
-    "train_samples_per_second": 1.463,
-    "train_steps_per_second": 0.183
+    "train_samples_per_second": 1.443,
+    "train_steps_per_second": 0.18
 }

trainer_state.json

@@ -10,122 +10,122 @@
   "log_history": [
     {
       "epoch": 0.8,
-      "grad_norm": 3.535655975341797,
+      "grad_norm": 3.2206196784973145,
       "learning_rate": 0.0001,
       "loss": 2.8223,
       "step": 1
     },
     {
       "epoch": 1.0,
-      "grad_norm": 4.1554155349731445,
+      "grad_norm": 3.775913715362549,
       "learning_rate": 0.0002,
       "loss": 2.8268,
       "step": 2
     },
     {
       "epoch": 1.0,
-      "eval_loss": 2.803156852722168,
-      "eval_runtime": 0.5136,
-      "eval_samples_per_second": 1.947,
-      "eval_steps_per_second": 1.947,
+      "eval_loss": 2.808971643447876,
+      "eval_runtime": 0.5089,
+      "eval_samples_per_second": 1.965,
+      "eval_steps_per_second": 1.965,
       "step": 2
     },
     {
       "epoch": 1.8,
-      "grad_norm": 2.6138558387756348,
+      "grad_norm": 2.4820940494537354,
       "learning_rate": 0.000175,
-      "loss": 2.6803,
+      "loss": 2.6863,
       "step": 3
     },
     {
       "epoch": 2.0,
-      "grad_norm": 2.1446619033813477,
+      "grad_norm": 2.1421539783477783,
       "learning_rate": 0.00015000000000000001,
-      "loss": 2.4443,
+      "loss": 2.4484,
       "step": 4
     },
     {
       "epoch": 2.0,
-      "eval_loss": 2.5913567543029785,
-      "eval_runtime": 0.5175,
-      "eval_samples_per_second": 1.932,
-      "eval_steps_per_second": 1.932,
+      "eval_loss": 2.5902364253997803,
+      "eval_runtime": 0.5114,
+      "eval_samples_per_second": 1.955,
+      "eval_steps_per_second": 1.955,
       "step": 4
     },
     {
       "epoch": 2.8,
-      "grad_norm": 1.7912520170211792,
+      "grad_norm": 1.6769822835922241,
       "learning_rate": 0.000125,
-      "loss": 2.4006,
+      "loss": 2.3966,
       "step": 5
     },
     {
       "epoch": 3.0,
-      "grad_norm": 2.2134900093078613,
+      "grad_norm": 2.2351601123809814,
       "learning_rate": 0.0001,
-      "loss": 2.3568,
+      "loss": 2.3563,
       "step": 6
     },
     {
       "epoch": 3.0,
-      "eval_loss": 2.456364870071411,
-      "eval_runtime": 0.5073,
-      "eval_samples_per_second": 1.971,
-      "eval_steps_per_second": 1.971,
+      "eval_loss": 2.461848020553589,
+      "eval_runtime": 0.5161,
+      "eval_samples_per_second": 1.938,
+      "eval_steps_per_second": 1.938,
       "step": 6
     },
     {
       "epoch": 3.8,
-      "grad_norm": 1.7705835103988647,
+      "grad_norm": 1.7321960926055908,
       "learning_rate": 7.500000000000001e-05,
-      "loss": 2.2664,
+      "loss": 2.2687,
       "step": 7
     },
     {
       "epoch": 4.0,
-      "grad_norm": 1.7635867595672607,
+      "grad_norm": 1.719794750213623,
       "learning_rate": 5e-05,
-      "loss": 2.1515,
+      "loss": 2.1538,
       "step": 8
     },
     {
       "epoch": 4.0,
-      "eval_loss": 2.3617074489593506,
-      "eval_runtime": 0.5027,
-      "eval_samples_per_second": 1.989,
-      "eval_steps_per_second": 1.989,
+      "eval_loss": 2.367581844329834,
+      "eval_runtime": 0.5201,
+      "eval_samples_per_second": 1.923,
+      "eval_steps_per_second": 1.923,
       "step": 8
     },
     {
       "epoch": 4.8,
-      "grad_norm": 1.2085267305374146,
+      "grad_norm": 1.1918765306472778,
       "learning_rate": 2.5e-05,
-      "loss": 2.1536,
+      "loss": 2.1569,
       "step": 9
     },
     {
       "epoch": 5.0,
-      "grad_norm": 1.5746992826461792,
+      "grad_norm": 1.5512747764587402,
       "learning_rate": 0.0,
-      "loss": 2.1324,
+      "loss": 2.1331,
       "step": 10
     },
     {
       "epoch": 5.0,
-      "eval_loss": 2.3257529735565186,
-      "eval_runtime": 0.5048,
-      "eval_samples_per_second": 1.981,
-      "eval_steps_per_second": 1.981,
+      "eval_loss": 2.331345319747925,
+      "eval_runtime": 0.5274,
+      "eval_samples_per_second": 1.896,
+      "eval_steps_per_second": 1.896,
       "step": 10
     },
     {
       "epoch": 5.0,
       "step": 10,
       "total_flos": 549306960445440.0,
-      "train_loss": 2.4234995365142824,
-      "train_runtime": 54.6732,
-      "train_samples_per_second": 1.463,
-      "train_steps_per_second": 0.183
+      "train_loss": 2.424940276145935,
+      "train_runtime": 55.452,
+      "train_samples_per_second": 1.443,
+      "train_steps_per_second": 0.18
     }
   ],
   "logging_steps": 1,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2b6901fa5e2bec60b8edebb9df49acc61f121566b10b9726cffffd490adbbb3e
+oid sha256:41b5663d1c02ca7635555e1721c9bb900b71c353e8230e82502e623eefda763f
 size 5624