SARG-ai/test-atomizer-finetune

adapter_config.json

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

all_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 1.0,
     "total_flos": 0.0,
-    "train_loss": 0.8888574626296759,
-    "train_runtime": 208.0574,
-    "train_samples_per_second": 1.538,
-    "train_steps_per_second": 0.769
+    "train_loss": 1.770367290686678,
+    "train_runtime": 102.5984,
+    "train_samples_per_second": 1.579,
+    "train_steps_per_second": 0.789
 }

model-00001-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7d6a81f098367efd1fa02caeccef867e85c4c06793770d7273ad8485ceebcb1b
+oid sha256:9c994a40e131080d93cbef61e284b793761b451006a34720672440c80096ccc7
 size 4877660776

model-00002-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d4bc5eef1aa9e9121dab58003f1ac54181ee040651140792ca487389265e52ab
+oid sha256:c7a96313da3d2879efc2fe61ac39e0cbe81b54452ab8013ae6746236840a7fb0
 size 4932751008

model-00003-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:20f4fdaca40f60e1f5e0bcebc90462f8a04dd27f8f76b392b00ecd9960d65da5
+oid sha256:43bfd5cad7d2fabaa3e01cc208c1e888fe999b67c9fa4a2c1ed8c57eedfabc0c
 size 4330865200

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 1.0,
     "total_flos": 0.0,
-    "train_loss": 0.8888574626296759,
-    "train_runtime": 208.0574,
-    "train_samples_per_second": 1.538,
-    "train_steps_per_second": 0.769
+    "train_loss": 1.770367290686678,
+    "train_runtime": 102.5984,
+    "train_samples_per_second": 1.579,
+    "train_steps_per_second": 0.789
 }

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": 13.950390815734863,
-      "learning_rate": 9.375e-05,
-      "loss": 8.2401,
+      "epoch": 0.12345679012345678,
+      "grad_norm": 12.38063907623291,
+      "learning_rate": 8.765432098765433e-05,
+      "loss": 7.922,
       "step": 10
     },
     {
-      "epoch": 0.125,
-      "grad_norm": 11.687304496765137,
-      "learning_rate": 8.75e-05,
-      "loss": 4.5168,
+      "epoch": 0.24691358024691357,
+      "grad_norm": 13.536417007446289,
+      "learning_rate": 7.530864197530865e-05,
+      "loss": 4.6077,
       "step": 20
     },
     {
-      "epoch": 0.1875,
-      "grad_norm": 0.31002017855644226,
-      "learning_rate": 8.125000000000001e-05,
-      "loss": 0.8313,
+      "epoch": 0.37037037037037035,
+      "grad_norm": 2.8598146438598633,
+      "learning_rate": 6.296296296296296e-05,
+      "loss": 1.327,
       "step": 30
     },
     {
-      "epoch": 0.25,
-      "grad_norm": 0.2928928732872009,
-      "learning_rate": 7.500000000000001e-05,
-      "loss": 0.0523,
+      "epoch": 0.49382716049382713,
+      "grad_norm": 0.32510578632354736,
+      "learning_rate": 5.061728395061729e-05,
+      "loss": 0.0959,
       "step": 40
     },
     {
-      "epoch": 0.3125,
-      "grad_norm": 0.3612825274467468,
-      "learning_rate": 6.875e-05,
-      "loss": 0.056,
+      "epoch": 0.6172839506172839,
+      "grad_norm": 0.2507149875164032,
+      "learning_rate": 3.82716049382716e-05,
+      "loss": 0.1085,
       "step": 50
     },
     {
-      "epoch": 0.375,
-      "grad_norm": 0.20652379095554352,
-      "learning_rate": 6.25e-05,
-      "loss": 0.0465,
+      "epoch": 0.7407407407407407,
+      "grad_norm": 0.2675971984863281,
+      "learning_rate": 2.5925925925925925e-05,
+      "loss": 0.0834,
       "step": 60
     },
     {
-      "epoch": 0.4375,
-      "grad_norm": 0.2031383514404297,
-      "learning_rate": 5.6250000000000005e-05,
-      "loss": 0.0512,
+      "epoch": 0.8641975308641975,
+      "grad_norm": 0.22273920476436615,
+      "learning_rate": 1.3580246913580247e-05,
+      "loss": 0.1037,
       "step": 70
     },
     {
-      "epoch": 0.5,
-      "grad_norm": 0.16576522588729858,
-      "learning_rate": 5e-05,
-      "loss": 0.0508,
+      "epoch": 0.9876543209876543,
+      "grad_norm": 0.20021985471248627,
+      "learning_rate": 1.234567901234568e-06,
+      "loss": 0.0836,
       "step": 80
     },
-    {
-      "epoch": 0.5625,
-      "grad_norm": 0.21390429139137268,
-      "learning_rate": 4.375e-05,
-      "loss": 0.0481,
-      "step": 90
-    },
-    {
-      "epoch": 0.5625,
-      "eval_runtime": 16.2003,
-      "eval_samples_per_second": 4.938,
-      "eval_steps_per_second": 0.309,
-      "step": 90
-    },
-    {
-      "epoch": 0.625,
-      "grad_norm": 0.2153244912624359,
-      "learning_rate": 3.7500000000000003e-05,
-      "loss": 0.0485,
-      "step": 100
-    },
-    {
-      "epoch": 0.6875,
-      "grad_norm": 0.1978517323732376,
-      "learning_rate": 3.125e-05,
-      "loss": 0.0485,
-      "step": 110
-    },
-    {
-      "epoch": 0.75,
-      "grad_norm": 0.24245601892471313,
-      "learning_rate": 2.5e-05,
-      "loss": 0.0452,
-      "step": 120
-    },
-    {
-      "epoch": 0.8125,
-      "grad_norm": 0.19363097846508026,
-      "learning_rate": 1.8750000000000002e-05,
-      "loss": 0.0462,
-      "step": 130
-    },
-    {
-      "epoch": 0.875,
-      "grad_norm": 0.17478688061237335,
-      "learning_rate": 1.25e-05,
-      "loss": 0.0527,
-      "step": 140
-    },
-    {
-      "epoch": 0.9375,
-      "grad_norm": 0.17106127738952637,
-      "learning_rate": 6.25e-06,
-      "loss": 0.045,
-      "step": 150
-    },
-    {
-      "epoch": 1.0,
-      "grad_norm": 0.1665540337562561,
-      "learning_rate": 0.0,
-      "loss": 0.0425,
-      "step": 160
-    },
     {
       "epoch": 1.0,
-      "eval_runtime": 14.9897,
-      "eval_samples_per_second": 5.337,
-      "eval_steps_per_second": 0.334,
-      "step": 160
+      "eval_runtime": 8.9544,
+      "eval_samples_per_second": 4.579,
+      "eval_steps_per_second": 0.335,
+      "step": 81
     },
     {
       "epoch": 1.0,
-      "step": 160,
+      "step": 81,
       "total_flos": 0.0,
-      "train_loss": 0.8888574626296759,
-      "train_runtime": 208.0574,
-      "train_samples_per_second": 1.538,
-      "train_steps_per_second": 0.769
+      "train_loss": 1.770367290686678,
+      "train_runtime": 102.5984,
+      "train_samples_per_second": 1.579,
+      "train_steps_per_second": 0.789
     }
   ],
   "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 sha256:76f2b98bc050288ce5addc969efa1900c5f5d81da51dc8a1a2bde72fe07d577d
+oid sha256:551092e3a36216a13e25d403df5e47bdf244b4e4c6981cf6cd7add30f473525c
 size 5432