Model save

README.md

@@ -2,15 +2,12 @@
 license: apache-2.0
 base_model: mistralai/Mistral-7B-Instruct-v0.2
 tags:
-- alignment-handbook
-- trl
-- sft
-- generated_from_trainer
 - trl
 - sft
+- alignment-handbook
 - generated_from_trainer
 datasets:
-- arcee-ai/Zilo-Filtered-SQL-Instruct
+- generator
 model-index:
 - name: zilo-instruct-v2-sft-filtered
   results: []
@@ -21,9 +18,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # zilo-instruct-v2-sft-filtered
 
-This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.2](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2) on the arcee-ai/Zilo-Filtered-SQL-Instruct dataset.
+This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.2](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2) on the generator dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.4110
+- Loss: 0.4709
 
 ## Model description
 
@@ -53,15 +50,17 @@ The following hyperparameters were used during training:
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: cosine
 - lr_scheduler_warmup_ratio: 0.1
-- num_epochs: 3
+- num_epochs: 5
 
 ### Training results
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 0.612         | 1.0   | 15   | 0.5663          |
-| 0.4089        | 2.0   | 30   | 0.4292          |
-| 0.3032        | 3.0   | 45   | 0.4110          |
+| 0.6543        | 1.0   | 15   | 0.5526          |
+| 0.4164        | 2.0   | 30   | 0.4357          |
+| 0.2956        | 3.0   | 45   | 0.4067          |
+| 0.1871        | 4.0   | 60   | 0.4253          |
+| 0.1363        | 5.0   | 75   | 0.4709          |
 
 
 ### Framework versions

all_results.json

@@ -1,14 +1,14 @@
 {
-    "epoch": 3.0,
+    "epoch": 5.0,
     "eval_loss": 0.4109707474708557,
     "eval_runtime": 4.066,
     "eval_samples": 2958,
     "eval_samples_per_second": 48.942,
     "eval_steps_per_second": 0.984,
-    "total_flos": 18844169011200.0,
-    "train_loss": 0.5794603930579292,
-    "train_runtime": 365.3228,
+    "total_flos": 31406948352000.0,
+    "train_loss": 0.4169128179550171,
+    "train_runtime": 607.0886,
     "train_samples": 26621,
-    "train_samples_per_second": 14.765,
-    "train_steps_per_second": 0.123
+    "train_samples_per_second": 14.808,
+    "train_steps_per_second": 0.124
 }

config.json

@@ -21,6 +21,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.41.1",
-  "use_cache": true,
+  "use_cache": false,
   "vocab_size": 32000
 }

model-00001-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7f405459d29d4f24ca925cc753af33355908ad63f30f4385a2e4a85f9d3f7479
+oid sha256:fc6da614ec17ae16d1e4117583f2468dde398379987e0368a6a7f30127ec8b8f
 size 4943162336

model-00002-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a2447be147c0ea3ec8d8acffff86e58e7449d6ef2067e5fe061246f32d036885
+oid sha256:afde27c7acd99cf421df092d50c893e786d8ff5bdcc984b8a89595cb88ca8ada
 size 4999819336

model-00003-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9009683f866f59b1167a0e3d05306e477ba5bbb876214564f0458cc1e2d653d4
+oid sha256:6f1e80481537111e6e033c6d7390504f3d0769dde8d969cb2f7599a32f7b1c64
 size 4540516344

runs/May24_22-46-04_5c31577a2818/events.out.tfevents.1716590804.5c31577a2818.6596.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6a11d7ec9e457fb408aab58496acf88e3a86df481d5a61e958ad7811656e9964
+size 6267

runs/May24_22-50-13_5c31577a2818/events.out.tfevents.1716591025.5c31577a2818.7690.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:83985bee703a1a9564efc903e6385fa47698b1e1a066a72aba84dbe0eb10bcc0
+size 8182

runs/May24_22-50-13_5c31577a2818/events.out.tfevents.1716591543.5c31577a2818.7690.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1eaef88f2128a6d15206f9f6ed81e30a94f464f7eac30ebd23dd4a0809832280
+size 354

runs/May24_23-58-51_5c31577a2818/events.out.tfevents.1716595145.5c31577a2818.11341.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7cd90bece5d909e72588e90ae6f1b13df19a4deeb4221c70a4738e57711a766
+size 9965

train_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 3.0,
-    "total_flos": 18844169011200.0,
-    "train_loss": 0.5794603930579292,
-    "train_runtime": 365.3228,
+    "epoch": 5.0,
+    "total_flos": 31406948352000.0,
+    "train_loss": 0.4169128179550171,
+    "train_runtime": 607.0886,
     "train_samples": 26621,
-    "train_samples_per_second": 14.765,
-    "train_steps_per_second": 0.123
+    "train_samples_per_second": 14.808,
+    "train_steps_per_second": 0.124
 }

trainer_state.json

@@ -1,121 +1,179 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 3.0,
+  "epoch": 5.0,
   "eval_steps": 500,
-  "global_step": 45,
+  "global_step": 75,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.06666666666666667,
-      "grad_norm": 98.21532030503785,
-      "learning_rate": 4.000000000000001e-06,
+      "grad_norm": 98.20704487515755,
+      "learning_rate": 2.5e-06,
       "loss": 1.8595,
       "step": 1
     },
     {
       "epoch": 0.3333333333333333,
-      "grad_norm": 30.800740243895696,
-      "learning_rate": 2e-05,
-      "loss": 1.3856,
+      "grad_norm": 26.158158491928763,
+      "learning_rate": 1.25e-05,
+      "loss": 1.3331,
       "step": 5
     },
     {
       "epoch": 0.6666666666666666,
-      "grad_norm": 15.616799282095581,
-      "learning_rate": 1.9238795325112867e-05,
-      "loss": 0.8256,
+      "grad_norm": 52.110102576445215,
+      "learning_rate": 1.9956059820218982e-05,
+      "loss": 0.7898,
       "step": 10
     },
     {
       "epoch": 1.0,
-      "grad_norm": 4.893885987624207,
-      "learning_rate": 1.7071067811865477e-05,
-      "loss": 0.612,
+      "grad_norm": 7.630410550452039,
+      "learning_rate": 1.9466156752904344e-05,
+      "loss": 0.6543,
       "step": 15
     },
     {
       "epoch": 1.0,
-      "eval_loss": 0.566252589225769,
-      "eval_runtime": 4.2539,
-      "eval_samples_per_second": 46.78,
+      "eval_loss": 0.552646815776825,
+      "eval_runtime": 4.2535,
+      "eval_samples_per_second": 46.785,
       "eval_steps_per_second": 0.94,
       "step": 15
     },
     {
       "epoch": 1.3333333333333333,
-      "grad_norm": 3.3286961705391014,
-      "learning_rate": 1.3826834323650899e-05,
-      "loss": 0.5009,
+      "grad_norm": 17.50115629098683,
+      "learning_rate": 1.8458320592590976e-05,
+      "loss": 0.5122,
       "step": 20
     },
     {
       "epoch": 1.6666666666666665,
-      "grad_norm": 1.3920639558462544,
-      "learning_rate": 1e-05,
-      "loss": 0.4407,
+      "grad_norm": 3.153164672454271,
+      "learning_rate": 1.698769427778842e-05,
+      "loss": 0.4774,
       "step": 25
     },
     {
       "epoch": 2.0,
-      "grad_norm": 1.3655223364170395,
-      "learning_rate": 6.173165676349103e-06,
-      "loss": 0.4089,
+      "grad_norm": 1.7735518447317222,
+      "learning_rate": 1.513474193514842e-05,
+      "loss": 0.4164,
       "step": 30
     },
     {
       "epoch": 2.0,
-      "eval_loss": 0.429166316986084,
-      "eval_runtime": 4.1164,
-      "eval_samples_per_second": 48.343,
-      "eval_steps_per_second": 0.972,
+      "eval_loss": 0.4356851875782013,
+      "eval_runtime": 4.105,
+      "eval_samples_per_second": 48.477,
+      "eval_steps_per_second": 0.974,
       "step": 30
     },
     {
       "epoch": 2.3333333333333335,
-      "grad_norm": 1.2577240969960297,
-      "learning_rate": 2.9289321881345257e-06,
-      "loss": 0.3299,
+      "grad_norm": 1.295932258196787,
+      "learning_rate": 1.300084635000341e-05,
+      "loss": 0.3318,
       "step": 35
     },
     {
       "epoch": 2.6666666666666665,
-      "grad_norm": 0.9884829876056015,
-      "learning_rate": 7.612046748871327e-07,
-      "loss": 0.3136,
+      "grad_norm": 1.0485408353072447,
+      "learning_rate": 1.070276188945293e-05,
+      "loss": 0.3158,
       "step": 40
     },
     {
       "epoch": 3.0,
-      "grad_norm": 1.4077416282701332,
-      "learning_rate": 0.0,
-      "loss": 0.3032,
+      "grad_norm": 1.4216337627274278,
+      "learning_rate": 8.366226381814698e-06,
+      "loss": 0.2956,
       "step": 45
     },
     {
       "epoch": 3.0,
-      "eval_loss": 0.4109707474708557,
-      "eval_runtime": 4.1074,
-      "eval_samples_per_second": 48.449,
-      "eval_steps_per_second": 0.974,
+      "eval_loss": 0.4067309498786926,
+      "eval_runtime": 4.1095,
+      "eval_samples_per_second": 48.424,
+      "eval_steps_per_second": 0.973,
       "step": 45
     },
     {
-      "epoch": 3.0,
-      "step": 45,
-      "total_flos": 18844169011200.0,
-      "train_loss": 0.5794603930579292,
-      "train_runtime": 365.3228,
-      "train_samples_per_second": 14.765,
-      "train_steps_per_second": 0.123
+      "epoch": 3.3333333333333335,
+      "grad_norm": 1.290531056985457,
+      "learning_rate": 6.119081473277502e-06,
+      "loss": 0.2149,
+      "step": 50
+    },
+    {
+      "epoch": 3.6666666666666665,
+      "grad_norm": 1.1012108900548532,
+      "learning_rate": 4.084277875864776e-06,
+      "loss": 0.1992,
+      "step": 55
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 1.1454117876381236,
+      "learning_rate": 2.373148218896182e-06,
+      "loss": 0.1871,
+      "step": 60
+    },
+    {
+      "epoch": 4.0,
+      "eval_loss": 0.42530176043510437,
+      "eval_runtime": 4.1023,
+      "eval_samples_per_second": 48.509,
+      "eval_steps_per_second": 0.975,
+      "step": 60
+    },
+    {
+      "epoch": 4.333333333333333,
+      "grad_norm": 0.947060255649263,
+      "learning_rate": 1.0793155744261352e-06,
+      "loss": 0.145,
+      "step": 65
+    },
+    {
+      "epoch": 4.666666666666667,
+      "grad_norm": 1.2233089117487779,
+      "learning_rate": 2.735709467518699e-07,
+      "loss": 0.1396,
+      "step": 70
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 0.8371390612015485,
+      "learning_rate": 0.0,
+      "loss": 0.1363,
+      "step": 75
+    },
+    {
+      "epoch": 5.0,
+      "eval_loss": 0.4708780348300934,
+      "eval_runtime": 4.0889,
+      "eval_samples_per_second": 48.669,
+      "eval_steps_per_second": 0.978,
+      "step": 75
+    },
+    {
+      "epoch": 5.0,
+      "step": 75,
+      "total_flos": 31406948352000.0,
+      "train_loss": 0.4169128179550171,
+      "train_runtime": 607.0886,
+      "train_samples_per_second": 14.808,
+      "train_steps_per_second": 0.124
     }
   ],
   "logging_steps": 5,
-  "max_steps": 45,
+  "max_steps": 75,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 3,
+  "num_train_epochs": 5,
   "save_steps": 100,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -129,7 +187,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 18844169011200.0,
+  "total_flos": 31406948352000.0,
   "train_batch_size": 16,
   "trial_name": null,
   "trial_params": null

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e6c1175d266c101e9bb917f7cfdadd36783940e29fb5163f64c8a645a0b647ca
-size 6328
+oid sha256:dcbd9b7605e96941aae059351fb08ef1d90212437053df7ce4e1a3a85ba81b73
+size 6392