Model save

README.md

@@ -2,16 +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: []
@@ -22,9 +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: 0.4072
+- Loss: 0.5474
 
 ## Model description
 
@@ -60,9 +56,9 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 0.6043        | 1.0   | 15   | 0.5420          |
-| 0.3922        | 2.0   | 30   | 0.4157          |
-| 0.2791        | 3.0   | 45   | 0.4072          |
+| 1.8602        | 1.0   | 7    | 1.3280          |
+| 2.9102        | 2.0   | 14   | 0.6256          |
+| 0.5329        | 3.0   | 21   | 0.5474          |
 
 
 ### Framework versions

all_results.json

@@ -5,10 +5,10 @@
     "eval_samples": 2958,
     "eval_samples_per_second": 49.046,
     "eval_steps_per_second": 0.986,
-    "total_flos": 18844169011200.0,
-    "train_loss": 0.5649417373869154,
-    "train_runtime": 363.236,
-    "train_samples": 26621,
-    "train_samples_per_second": 14.85,
-    "train_steps_per_second": 0.124
+    "total_flos": 8793945538560.0,
+    "train_loss": 1.4474536804925828,
+    "train_runtime": 170.8306,
+    "train_samples": 12338,
+    "train_samples_per_second": 13.821,
+    "train_steps_per_second": 0.123
 }

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:ce24b378ae43fe02b92e39dd4d0181fc4821e81b76e1ab3b48feb485cee8ad9a
+oid sha256:8bdc26479b269dae211f4fb563e33dc9a9e321a8c077d99b7dd886866e5a0497
 size 4943162336

model-00002-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:767652637ec5730e9771e3f2929b36cf446fdb16382700a907dde817a6c9b649
+oid sha256:b6b5bd3f865efc671963fb864ab3b68c1a96cef2e962caa42aa0f5468d01239b
 size 4999819336

model-00003-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:90277adea4cd1317fb6624f070f075f16342cb8b18b00043f18cb5db49b56010
+oid sha256:27f029b043e2a1a93f6f20165da76cb1cfa9516b9209497f4251e497cfe43580
 size 4540516344

runs/May25_01-35-59_5c31577a2818/events.out.tfevents.1716600985.5c31577a2818.16735.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55a671103226617f87cac9231aeffb3ef62914cd8bfdd7c80312ed1fe8973e97
+size 7156

runs/May25_01-40-07_5c31577a2818/events.out.tfevents.1716601219.5c31577a2818.17665.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8a6d92a72ffe3e45c7603637c9ac3a32d6af5afe2d2a3ca6330f71e60d3d47cc
+size 7156

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 3.0,
-    "total_flos": 18844169011200.0,
-    "train_loss": 0.5649417373869154,
-    "train_runtime": 363.236,
-    "train_samples": 26621,
-    "train_samples_per_second": 14.85,
-    "train_steps_per_second": 0.124
+    "total_flos": 8793945538560.0,
+    "train_loss": 1.4474536804925828,
+    "train_runtime": 170.8306,
+    "train_samples": 12338,
+    "train_samples_per_second": 13.821,
+    "train_steps_per_second": 0.123
 }

trainer_state.json

@@ -3,117 +3,82 @@
   "best_model_checkpoint": null,
   "epoch": 3.0,
   "eval_steps": 500,
-  "global_step": 45,
+  "global_step": 21,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.06666666666666667,
-      "grad_norm": 98.21472515172992,
-      "learning_rate": 4.000000000000001e-06,
-      "loss": 1.8595,
+      "epoch": 0.14285714285714285,
+      "grad_norm": 99.75653578103409,
+      "learning_rate": 6.666666666666667e-06,
+      "loss": 1.7828,
       "step": 1
     },
     {
-      "epoch": 0.3333333333333333,
-      "grad_norm": 30.799283162433554,
-      "learning_rate": 2e-05,
-      "loss": 1.3856,
+      "epoch": 0.7142857142857143,
+      "grad_norm": 421.35114809358015,
+      "learning_rate": 1.9396926207859085e-05,
+      "loss": 1.8602,
       "step": 5
     },
-    {
-      "epoch": 0.6666666666666666,
-      "grad_norm": 7.8774399080468624,
-      "learning_rate": 1.9238795325112867e-05,
-      "loss": 0.8251,
-      "step": 10
-    },
-    {
-      "epoch": 1.0,
-      "grad_norm": 4.085781086210361,
-      "learning_rate": 1.7071067811865477e-05,
-      "loss": 0.6043,
-      "step": 15
-    },
     {
       "epoch": 1.0,
-      "eval_loss": 0.5420283079147339,
-      "eval_runtime": 4.2361,
-      "eval_samples_per_second": 46.977,
-      "eval_steps_per_second": 0.944,
-      "step": 15
-    },
-    {
-      "epoch": 1.3333333333333333,
-      "grad_norm": 2.0713099775720534,
-      "learning_rate": 1.3826834323650899e-05,
-      "loss": 0.4794,
-      "step": 20
-    },
-    {
-      "epoch": 1.6666666666666665,
-      "grad_norm": 1.3136102966664787,
-      "learning_rate": 1e-05,
-      "loss": 0.4248,
-      "step": 25
+      "eval_loss": 1.3279948234558105,
+      "eval_runtime": 2.2021,
+      "eval_samples_per_second": 39.054,
+      "eval_steps_per_second": 0.908,
+      "step": 7
     },
     {
-      "epoch": 2.0,
-      "grad_norm": 1.1912342263078068,
-      "learning_rate": 6.173165676349103e-06,
-      "loss": 0.3922,
-      "step": 30
+      "epoch": 1.4285714285714286,
+      "grad_norm": 60.4301410351058,
+      "learning_rate": 1.342020143325669e-05,
+      "loss": 2.9102,
+      "step": 10
     },
     {
       "epoch": 2.0,
-      "eval_loss": 0.4156629145145416,
-      "eval_runtime": 4.0917,
-      "eval_samples_per_second": 48.635,
-      "eval_steps_per_second": 0.978,
-      "step": 30
+      "eval_loss": 0.6256123781204224,
+      "eval_runtime": 2.0607,
+      "eval_samples_per_second": 41.733,
+      "eval_steps_per_second": 0.971,
+      "step": 14
     },
     {
-      "epoch": 2.3333333333333335,
-      "grad_norm": 1.1730848362574522,
-      "learning_rate": 2.9289321881345257e-06,
-      "loss": 0.3075,
-      "step": 35
-    },
-    {
-      "epoch": 2.6666666666666665,
-      "grad_norm": 0.9598398180975349,
-      "learning_rate": 7.612046748871327e-07,
-      "loss": 0.2917,
-      "step": 40
+      "epoch": 2.142857142857143,
+      "grad_norm": 9.507888146594869,
+      "learning_rate": 5.000000000000003e-06,
+      "loss": 0.686,
+      "step": 15
     },
     {
-      "epoch": 3.0,
-      "grad_norm": 1.3661284541735195,
-      "learning_rate": 0.0,
-      "loss": 0.2791,
-      "step": 45
+      "epoch": 2.857142857142857,
+      "grad_norm": 1.2620367321960468,
+      "learning_rate": 1.519224698779198e-07,
+      "loss": 0.5329,
+      "step": 20
     },
     {
       "epoch": 3.0,
-      "eval_loss": 0.40715718269348145,
-      "eval_runtime": 4.0987,
-      "eval_samples_per_second": 48.552,
-      "eval_steps_per_second": 0.976,
-      "step": 45
+      "eval_loss": 0.5474353432655334,
+      "eval_runtime": 2.0545,
+      "eval_samples_per_second": 41.86,
+      "eval_steps_per_second": 0.973,
+      "step": 21
     },
     {
       "epoch": 3.0,
-      "step": 45,
-      "total_flos": 18844169011200.0,
-      "train_loss": 0.5649417373869154,
-      "train_runtime": 363.236,
-      "train_samples_per_second": 14.85,
-      "train_steps_per_second": 0.124
+      "step": 21,
+      "total_flos": 8793945538560.0,
+      "train_loss": 1.4474536804925828,
+      "train_runtime": 170.8306,
+      "train_samples_per_second": 13.821,
+      "train_steps_per_second": 0.123
     }
   ],
   "logging_steps": 5,
-  "max_steps": 45,
+  "max_steps": 21,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 3,
   "save_steps": 100,
@@ -129,7 +94,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 18844169011200.0,
+  "total_flos": 8793945538560.0,
   "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 sha256:16468e20b0c8c5a6af8b75ded9d347ce0856a32d5e19a9f872bac737dba7ae17
+oid sha256:c16c304b54e234c167d4a62517da9fcb81b63bf76374bb1ffd8fd038494174c9
 size 6392