SARG-ai/test-atomizer-finetune

adapter_config.json

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

all_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 1.0,
     "total_flos": 0.0,
-    "train_loss": 0.8959081767126917,
-    "train_runtime": 208.3433,
-    "train_samples_per_second": 1.536,
-    "train_steps_per_second": 0.768
+    "train_loss": 1.7328798631091176,
+    "train_runtime": 102.349,
+    "train_samples_per_second": 1.583,
+    "train_steps_per_second": 0.791
 }

model-00001-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ad44180e7c11905334ab8dd4ea8df8264b9b7b6754251b75f974fb350be6ff89
+oid sha256:d12cef8c52291ab9dd85aaa3ff0b585a43112ab5ab7b4391f761aa57fb32fa8b
 size 4877660776

model-00002-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9bbc2954dbeb5e39b9a77d78eae9895967c2a5b487cf74285eae2b645df6ea10
+oid sha256:f094a4f0c621e5c2d0774071594f20de88271fbf57f8794f3c96842df16a6715
 size 4932751008

model-00003-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0c710f806e8c0dd2cbc2d1a28fce2ee8c45ae9713680333a65d1918e0d14b614
+oid sha256:7017812696b56dca73da804070c2483535df82cb51617625a5ccd457b514349d
 size 4330865200

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 1.0,
     "total_flos": 0.0,
-    "train_loss": 0.8959081767126917,
-    "train_runtime": 208.3433,
-    "train_samples_per_second": 1.536,
-    "train_steps_per_second": 0.768
+    "train_loss": 1.7328798631091176,
+    "train_runtime": 102.349,
+    "train_samples_per_second": 1.583,
+    "train_steps_per_second": 0.791
 }

trainer_state.json

@@ -3,149 +3,86 @@
   "best_model_checkpoint": null,
   "epoch": 1.0,
   "eval_steps": 500,
-  "global_step": 160,
+  "global_step": 81,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.0625,
-      "grad_norm": 12.499859809875488,
-      "learning_rate": 9.375e-05,
-      "loss": 8.224,
+      "epoch": 0.12345679012345678,
+      "grad_norm": 13.576780319213867,
+      "learning_rate": 8.765432098765433e-05,
+      "loss": 7.8907,
       "step": 10
     },
     {
-      "epoch": 0.125,
-      "grad_norm": 11.91177749633789,
-      "learning_rate": 8.75e-05,
-      "loss": 4.5631,
+      "epoch": 0.24691358024691357,
+      "grad_norm": 11.127266883850098,
+      "learning_rate": 7.530864197530865e-05,
+      "loss": 4.5017,
       "step": 20
     },
     {
-      "epoch": 0.1875,
-      "grad_norm": 0.274835467338562,
-      "learning_rate": 8.125000000000001e-05,
-      "loss": 0.8739,
+      "epoch": 0.37037037037037035,
+      "grad_norm": 1.6644705533981323,
+      "learning_rate": 6.296296296296296e-05,
+      "loss": 1.1627,
       "step": 30
     },
     {
-      "epoch": 0.25,
-      "grad_norm": 0.2978007197380066,
-      "learning_rate": 7.500000000000001e-05,
-      "loss": 0.0493,
+      "epoch": 0.49382716049382713,
+      "grad_norm": 0.319928914308548,
+      "learning_rate": 5.061728395061729e-05,
+      "loss": 0.094,
       "step": 40
     },
     {
-      "epoch": 0.3125,
-      "grad_norm": 0.6643026471138,
-      "learning_rate": 6.875e-05,
-      "loss": 0.0554,
+      "epoch": 0.6172839506172839,
+      "grad_norm": 0.2718435227870941,
+      "learning_rate": 3.82716049382716e-05,
+      "loss": 0.1081,
       "step": 50
     },
     {
-      "epoch": 0.375,
-      "grad_norm": 0.41495800018310547,
-      "learning_rate": 6.25e-05,
-      "loss": 0.087,
+      "epoch": 0.7407407407407407,
+      "grad_norm": 0.291215717792511,
+      "learning_rate": 2.5925925925925925e-05,
+      "loss": 0.0831,
       "step": 60
     },
     {
-      "epoch": 0.4375,
-      "grad_norm": 0.1979512870311737,
-      "learning_rate": 5.6250000000000005e-05,
-      "loss": 0.0535,
+      "epoch": 0.8641975308641975,
+      "grad_norm": 0.232026606798172,
+      "learning_rate": 1.3580246913580247e-05,
+      "loss": 0.1037,
       "step": 70
     },
     {
-      "epoch": 0.5,
-      "grad_norm": 0.16608962416648865,
-      "learning_rate": 5e-05,
-      "loss": 0.0507,
+      "epoch": 0.9876543209876543,
+      "grad_norm": 0.20218397676944733,
+      "learning_rate": 1.234567901234568e-06,
+      "loss": 0.0841,
       "step": 80
     },
-    {
-      "epoch": 0.5625,
-      "grad_norm": 0.23009341955184937,
-      "learning_rate": 4.375e-05,
-      "loss": 0.0481,
-      "step": 90
-    },
-    {
-      "epoch": 0.625,
-      "grad_norm": 0.20877023041248322,
-      "learning_rate": 3.7500000000000003e-05,
-      "loss": 0.0489,
-      "step": 100
-    },
-    {
-      "epoch": 0.6875,
-      "grad_norm": 0.2019713819026947,
-      "learning_rate": 3.125e-05,
-      "loss": 0.0483,
-      "step": 110
-    },
-    {
-      "epoch": 0.75,
-      "grad_norm": 0.24405285716056824,
-      "learning_rate": 2.5e-05,
-      "loss": 0.0453,
-      "step": 120
-    },
-    {
-      "epoch": 0.75,
-      "eval_runtime": 16.3044,
-      "eval_samples_per_second": 4.907,
-      "eval_steps_per_second": 0.307,
-      "step": 120
-    },
-    {
-      "epoch": 0.8125,
-      "grad_norm": 0.20071814954280853,
-      "learning_rate": 1.8750000000000002e-05,
-      "loss": 0.0464,
-      "step": 130
-    },
-    {
-      "epoch": 0.875,
-      "grad_norm": 0.17115017771720886,
-      "learning_rate": 1.25e-05,
-      "loss": 0.0529,
-      "step": 140
-    },
-    {
-      "epoch": 0.9375,
-      "grad_norm": 0.17328381538391113,
-      "learning_rate": 6.25e-06,
-      "loss": 0.0449,
-      "step": 150
-    },
-    {
-      "epoch": 1.0,
-      "grad_norm": 0.17866580188274384,
-      "learning_rate": 0.0,
-      "loss": 0.0427,
-      "step": 160
-    },
     {
       "epoch": 1.0,
-      "eval_runtime": 15.0976,
-      "eval_samples_per_second": 5.299,
-      "eval_steps_per_second": 0.331,
-      "step": 160
+      "eval_runtime": 8.9795,
+      "eval_samples_per_second": 4.566,
+      "eval_steps_per_second": 0.334,
+      "step": 81
     },
     {
       "epoch": 1.0,
-      "step": 160,
+      "step": 81,
       "total_flos": 0.0,
-      "train_loss": 0.8959081767126917,
-      "train_runtime": 208.3433,
-      "train_samples_per_second": 1.536,
-      "train_steps_per_second": 0.768
+      "train_loss": 1.7328798631091176,
+      "train_runtime": 102.349,
+      "train_samples_per_second": 1.583,
+      "train_steps_per_second": 0.791
     }
   ],
   "logging_steps": 10,
-  "max_steps": 160,
+  "max_steps": 81,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 1,
   "save_steps": 500,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:aa53b06a5d4b3333498d6359e0a5eecdfd27dcff325defbe2fadcecd4c971dfd
+oid sha256:1fef5bfe431ad9a004be5ad4894bf939cb99aaee0632ba9e885294a53a815d82
 size 5432