Model save

README.md

@@ -0,0 +1,68 @@
+---
+license: mit
+library_name: peft
+tags:
+- trl
+- sft
+- generated_from_trainer
+base_model: HuggingFaceH4/zephyr-7b-beta
+model-index:
+- name: skills_metaphor_chat-skills_red_herring_chat-lora
+  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. -->
+
+# skills_metaphor_chat-skills_red_herring_chat-lora
+
+This model is a fine-tuned version of [HuggingFaceH4/zephyr-7b-beta](https://huggingface.co/HuggingFaceH4/zephyr-7b-beta) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.2436
+
+## 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: 0.0002
+- train_batch_size: 4
+- eval_batch_size: 8
+- seed: 42
+- distributed_type: multi-GPU
+- gradient_accumulation_steps: 4
+- total_train_batch_size: 16
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: cosine
+- lr_scheduler_warmup_ratio: 0.1
+- num_epochs: 4.0
+
+### Training results
+
+| Training Loss | Epoch | Step | Validation Loss |
+|:-------------:|:-----:|:----:|:---------------:|
+| 0.2002        | 0.99  | 37   | 0.2199          |
+| 0.1723        | 2.0   | 75   | 0.2118          |
+| 0.1156        | 2.99  | 112  | 0.2252          |
+| 0.0804        | 3.95  | 148  | 0.2436          |
+
+
+### Framework versions
+
+- PEFT 0.7.1
+- Transformers 4.37.2
+- Pytorch 2.1.2+cu121
+- Datasets 2.14.6
+- Tokenizers 0.15.1

all_results.json

@@ -0,0 +1,13 @@
+{
+    "epoch": 3.95,
+    "eval_loss": 0.2436458021402359,
+    "eval_runtime": 7.9233,
+    "eval_samples": 200,
+    "eval_samples_per_second": 25.242,
+    "eval_steps_per_second": 3.155,
+    "train_loss": 0.2733316945063101,
+    "train_runtime": 797.9653,
+    "train_samples": 600,
+    "train_samples_per_second": 3.008,
+    "train_steps_per_second": 0.185
+}

eval_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 3.95,
+    "eval_loss": 0.2436458021402359,
+    "eval_runtime": 7.9233,
+    "eval_samples": 200,
+    "eval_samples_per_second": 25.242,
+    "eval_steps_per_second": 3.155
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 3.95,
+    "train_loss": 0.2733316945063101,
+    "train_runtime": 797.9653,
+    "train_samples": 600,
+    "train_samples_per_second": 3.008,
+    "train_steps_per_second": 0.185
+}

trainer_state.json

@@ -0,0 +1,242 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.9466666666666668,
+  "eval_steps": 500,
+  "global_step": 148,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.3333333333333333e-05,
+      "loss": 2.2275,
+      "step": 1
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 6.666666666666667e-05,
+      "loss": 2.2349,
+      "step": 5
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.00013333333333333334,
+      "loss": 1.4549,
+      "step": 10
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 0.0002,
+      "loss": 0.4894,
+      "step": 15
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.00019930337092856243,
+      "loss": 0.269,
+      "step": 20
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 0.00019722318955551306,
+      "loss": 0.2445,
+      "step": 25
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 0.00019378843817721854,
+      "loss": 0.2318,
+      "step": 30
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 0.00018904697174694447,
+      "loss": 0.2002,
+      "step": 35
+    },
+    {
+      "epoch": 0.99,
+      "eval_loss": 0.21994143724441528,
+      "eval_runtime": 8.8446,
+      "eval_samples_per_second": 22.613,
+      "eval_steps_per_second": 2.827,
+      "step": 37
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 0.0001830648511318223,
+      "loss": 0.2005,
+      "step": 40
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 0.00017592542271443887,
+      "loss": 0.1843,
+      "step": 45
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 0.00016772815716257412,
+      "loss": 0.1765,
+      "step": 50
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 0.00015858726354602248,
+      "loss": 0.1796,
+      "step": 55
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 0.00014863009810942815,
+      "loss": 0.1765,
+      "step": 60
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 0.000137995389871036,
+      "loss": 0.1685,
+      "step": 65
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 0.0001268313077693485,
+      "loss": 0.1616,
+      "step": 70
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 0.0001152933962873246,
+      "loss": 0.1723,
+      "step": 75
+    },
+    {
+      "epoch": 2.0,
+      "eval_loss": 0.21179074048995972,
+      "eval_runtime": 7.9847,
+      "eval_samples_per_second": 25.048,
+      "eval_steps_per_second": 3.131,
+      "step": 75
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 0.00010354240831620541,
+      "loss": 0.1375,
+      "step": 80
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 9.174206545276677e-05,
+      "loss": 0.1257,
+      "step": 85
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 8.005677693484077e-05,
+      "loss": 0.1262,
+      "step": 90
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 6.864934899622191e-05,
+      "loss": 0.1298,
+      "step": 95
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 5.767871655555751e-05,
+      "loss": 0.1154,
+      "step": 100
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 4.729772884265212e-05,
+      "loss": 0.12,
+      "step": 105
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 3.7651019814126654e-05,
+      "loss": 0.1156,
+      "step": 110
+    },
+    {
+      "epoch": 2.99,
+      "eval_loss": 0.2252231389284134,
+      "eval_runtime": 8.0616,
+      "eval_samples_per_second": 24.809,
+      "eval_steps_per_second": 3.101,
+      "step": 112
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 2.8872993029040508e-05,
+      "loss": 0.1084,
+      "step": 115
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 2.1085949060360654e-05,
+      "loss": 0.0907,
+      "step": 120
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 1.439838153227e-05,
+      "loss": 0.0856,
+      "step": 125
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 8.903465523913957e-06,
+      "loss": 0.087,
+      "step": 130
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 4.6777594000230855e-06,
+      "loss": 0.0818,
+      "step": 135
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 1.7801381552624563e-06,
+      "loss": 0.0932,
+      "step": 140
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 2.509731335744281e-07,
+      "loss": 0.0804,
+      "step": 145
+    },
+    {
+      "epoch": 3.95,
+      "eval_loss": 0.2436458021402359,
+      "eval_runtime": 7.8963,
+      "eval_samples_per_second": 25.328,
+      "eval_steps_per_second": 3.166,
+      "step": 148
+    },
+    {
+      "epoch": 3.95,
+      "step": 148,
+      "total_flos": 82630925025280.0,
+      "train_loss": 0.2733316945063101,
+      "train_runtime": 797.9653,
+      "train_samples_per_second": 3.008,
+      "train_steps_per_second": 0.185
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 148,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 4,
+  "save_steps": 500,
+  "total_flos": 82630925025280.0,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}