Model save

README.md

@@ -0,0 +1,76 @@
+---
+library_name: transformers
+license: apache-2.0
+base_model: alignment-handbook/zephyr-7b-sft-full
+tags:
+- trl
+- sft
+- generated_from_trainer
+datasets:
+- generator
+model-index:
+- name: zephyr-7b-stem-case-7
+  results: []
+---
+
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+# zephyr-7b-stem-case-7
+
+This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on the generator dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.0095
+
+## Model description
+
+More information needed
+
+## Intended uses & limitations
+
+More information needed
+
+## Training and evaluation data
+
+More information needed
+
+## Training procedure
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+- learning_rate: 1e-05
+- train_batch_size: 8
+- eval_batch_size: 8
+- seed: 42
+- distributed_type: multi-GPU
+- num_devices: 8
+- total_train_batch_size: 64
+- total_eval_batch_size: 64
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: cosine
+- lr_scheduler_warmup_ratio: 0.03
+- num_epochs: 10
+
+### Training results
+
+| Training Loss | Epoch | Step | Validation Loss |
+|:-------------:|:-----:|:----:|:---------------:|
+| 1.035         | 1.0   | 7    | 0.8066          |
+| 0.8612        | 2.0   | 14   | 0.5046          |
+| 0.4464        | 3.0   | 21   | 0.2836          |
+| 0.2872        | 4.0   | 28   | 0.1444          |
+| 0.1132        | 5.0   | 35   | 0.0621          |
+| 0.0555        | 6.0   | 42   | 0.0280          |
+| 0.0371        | 7.0   | 49   | 0.0172          |
+| 0.017         | 8.0   | 56   | 0.0122          |
+| 0.0126        | 9.0   | 63   | 0.0100          |
+| 0.0099        | 10.0  | 70   | 0.0095          |
+
+
+### Framework versions
+
+- Transformers 4.44.2
+- Pytorch 2.4.1+cu121
+- Datasets 3.0.0
+- Tokenizers 0.19.1

all_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 10.0,
+    "total_flos": 29313151795200.0,
+    "train_loss": 0.2696564427443913,
+    "train_runtime": 988.8828,
+    "train_samples": 2847,
+    "train_samples_per_second": 4.53,
+    "train_steps_per_second": 0.071
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "transformers_version": "4.44.2"
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 10.0,
+    "total_flos": 29313151795200.0,
+    "train_loss": 0.2696564427443913,
+    "train_runtime": 988.8828,
+    "train_samples": 2847,
+    "train_samples_per_second": 4.53,
+    "train_steps_per_second": 0.071
+}

trainer_state.json

@@ -0,0 +1,227 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 10.0,
+  "eval_steps": 500,
+  "global_step": 70,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.14285714285714285,
+      "grad_norm": 7.2481944316327,
+      "learning_rate": 3.3333333333333333e-06,
+      "loss": 1.1087,
+      "step": 1
+    },
+    {
+      "epoch": 0.7142857142857143,
+      "grad_norm": 6.492266623182951,
+      "learning_rate": 9.978029910109491e-06,
+      "loss": 1.035,
+      "step": 5
+    },
+    {
+      "epoch": 1.0,
+      "eval_loss": 0.8065708875656128,
+      "eval_runtime": 18.8479,
+      "eval_samples_per_second": 23.769,
+      "eval_steps_per_second": 0.371,
+      "step": 7
+    },
+    {
+      "epoch": 1.4285714285714286,
+      "grad_norm": 4.2798539792652965,
+      "learning_rate": 9.733078376452172e-06,
+      "loss": 0.8612,
+      "step": 10
+    },
+    {
+      "epoch": 2.0,
+      "eval_loss": 0.5046278834342957,
+      "eval_runtime": 18.7201,
+      "eval_samples_per_second": 23.931,
+      "eval_steps_per_second": 0.374,
+      "step": 14
+    },
+    {
+      "epoch": 2.142857142857143,
+      "grad_norm": 3.8618869981058155,
+      "learning_rate": 9.229160296295488e-06,
+      "loss": 0.6608,
+      "step": 15
+    },
+    {
+      "epoch": 2.857142857142857,
+      "grad_norm": 2.5003252439116106,
+      "learning_rate": 8.49384713889421e-06,
+      "loss": 0.4464,
+      "step": 20
+    },
+    {
+      "epoch": 3.0,
+      "eval_loss": 0.2836165726184845,
+      "eval_runtime": 18.7006,
+      "eval_samples_per_second": 23.956,
+      "eval_steps_per_second": 0.374,
+      "step": 21
+    },
+    {
+      "epoch": 3.571428571428571,
+      "grad_norm": 3.8783540297730106,
+      "learning_rate": 7.56737096757421e-06,
+      "loss": 0.2872,
+      "step": 25
+    },
+    {
+      "epoch": 4.0,
+      "eval_loss": 0.14440329372882843,
+      "eval_runtime": 18.7131,
+      "eval_samples_per_second": 23.94,
+      "eval_steps_per_second": 0.374,
+      "step": 28
+    },
+    {
+      "epoch": 4.285714285714286,
+      "grad_norm": 2.6373488856495295,
+      "learning_rate": 6.500423175001705e-06,
+      "loss": 0.1884,
+      "step": 30
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 2.524435640105693,
+      "learning_rate": 5.351380944726465e-06,
+      "loss": 0.1132,
+      "step": 35
+    },
+    {
+      "epoch": 5.0,
+      "eval_loss": 0.062061700969934464,
+      "eval_runtime": 18.7255,
+      "eval_samples_per_second": 23.925,
+      "eval_steps_per_second": 0.374,
+      "step": 35
+    },
+    {
+      "epoch": 5.714285714285714,
+      "grad_norm": 1.7870565823017404,
+      "learning_rate": 4.183113190907349e-06,
+      "loss": 0.0555,
+      "step": 40
+    },
+    {
+      "epoch": 6.0,
+      "eval_loss": 0.028033925220370293,
+      "eval_runtime": 18.7711,
+      "eval_samples_per_second": 23.866,
+      "eval_steps_per_second": 0.373,
+      "step": 42
+    },
+    {
+      "epoch": 6.428571428571429,
+      "grad_norm": 0.93279002933571,
+      "learning_rate": 3.059540736638751e-06,
+      "loss": 0.0371,
+      "step": 45
+    },
+    {
+      "epoch": 7.0,
+      "eval_loss": 0.017189698293805122,
+      "eval_runtime": 18.7331,
+      "eval_samples_per_second": 23.915,
+      "eval_steps_per_second": 0.374,
+      "step": 49
+    },
+    {
+      "epoch": 7.142857142857143,
+      "grad_norm": 0.6713728183393005,
+      "learning_rate": 2.042138937932388e-06,
+      "loss": 0.0254,
+      "step": 50
+    },
+    {
+      "epoch": 7.857142857142857,
+      "grad_norm": 0.6733311496948943,
+      "learning_rate": 1.186574109448091e-06,
+      "loss": 0.017,
+      "step": 55
+    },
+    {
+      "epoch": 8.0,
+      "eval_loss": 0.012202958576381207,
+      "eval_runtime": 18.6716,
+      "eval_samples_per_second": 23.994,
+      "eval_steps_per_second": 0.375,
+      "step": 56
+    },
+    {
+      "epoch": 8.571428571428571,
+      "grad_norm": 0.3925640628854087,
+      "learning_rate": 5.396577872130676e-07,
+      "loss": 0.0126,
+      "step": 60
+    },
+    {
+      "epoch": 9.0,
+      "eval_loss": 0.0099651413038373,
+      "eval_runtime": 18.7021,
+      "eval_samples_per_second": 23.955,
+      "eval_steps_per_second": 0.374,
+      "step": 63
+    },
+    {
+      "epoch": 9.285714285714286,
+      "grad_norm": 0.30666444604194676,
+      "learning_rate": 1.3678547337593494e-07,
+      "loss": 0.0107,
+      "step": 65
+    },
+    {
+      "epoch": 10.0,
+      "grad_norm": 0.3132195964128779,
+      "learning_rate": 0.0,
+      "loss": 0.0099,
+      "step": 70
+    },
+    {
+      "epoch": 10.0,
+      "eval_loss": 0.009540632367134094,
+      "eval_runtime": 18.6112,
+      "eval_samples_per_second": 24.071,
+      "eval_steps_per_second": 0.376,
+      "step": 70
+    },
+    {
+      "epoch": 10.0,
+      "step": 70,
+      "total_flos": 29313151795200.0,
+      "train_loss": 0.2696564427443913,
+      "train_runtime": 988.8828,
+      "train_samples_per_second": 4.53,
+      "train_steps_per_second": 0.071
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 70,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 100,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 29313151795200.0,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}