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_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_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: 0.2122
+
+## 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 |
+|:-------------:|:-----:|:----:|:---------------:|
+| 0.3285        | 0.96  | 18   | 0.2505          |
+| 0.1944        | 1.97  | 37   | 0.2189          |
+| 0.1767        | 2.99  | 56   | 0.2127          |
+| 0.1591        | 3.84  | 72   | 0.2122          |
+
+
+### 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": 3.84,
+    "eval_loss": 0.21221186220645905,
+    "eval_runtime": 4.0792,
+    "eval_samples": 100,
+    "eval_samples_per_second": 24.515,
+    "eval_steps_per_second": 3.187,
+    "train_loss": 0.4049788423710399,
+    "train_runtime": 399.0811,
+    "train_samples": 300,
+    "train_samples_per_second": 3.007,
+    "train_steps_per_second": 0.18
+}

eval_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 3.84,
+    "eval_loss": 0.21221186220645905,
+    "eval_runtime": 4.0792,
+    "eval_samples": 100,
+    "eval_samples_per_second": 24.515,
+    "eval_steps_per_second": 3.187
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 3.84,
+    "train_loss": 0.4049788423710399,
+    "train_runtime": 399.0811,
+    "train_samples": 300,
+    "train_samples_per_second": 3.007,
+    "train_steps_per_second": 0.18
+}

trainer_state.json

@@ -0,0 +1,152 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.84,
+  "eval_steps": 500,
+  "global_step": 72,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.5e-05,
+      "loss": 2.4616,
+      "step": 1
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 0.000125,
+      "loss": 2.2926,
+      "step": 5
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 0.0001995184726672197,
+      "loss": 1.0788,
+      "step": 10
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 0.00019415440651830208,
+      "loss": 0.3285,
+      "step": 15
+    },
+    {
+      "epoch": 0.96,
+      "eval_loss": 0.25046542286872864,
+      "eval_runtime": 4.9708,
+      "eval_samples_per_second": 20.118,
+      "eval_steps_per_second": 2.615,
+      "step": 18
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 0.00018314696123025454,
+      "loss": 0.2382,
+      "step": 20
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 0.00016715589548470185,
+      "loss": 0.2235,
+      "step": 25
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 0.0001471396736825998,
+      "loss": 0.2044,
+      "step": 30
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 0.0001242980179903264,
+      "loss": 0.1944,
+      "step": 35
+    },
+    {
+      "epoch": 1.97,
+      "eval_loss": 0.21891021728515625,
+      "eval_runtime": 4.0841,
+      "eval_samples_per_second": 24.485,
+      "eval_steps_per_second": 3.183,
+      "step": 37
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 0.0001,
+      "loss": 0.1881,
+      "step": 40
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 7.570198200967362e-05,
+      "loss": 0.1768,
+      "step": 45
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 5.286032631740023e-05,
+      "loss": 0.1628,
+      "step": 50
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 3.2844104515298155e-05,
+      "loss": 0.1767,
+      "step": 55
+    },
+    {
+      "epoch": 2.99,
+      "eval_loss": 0.2127072662115097,
+      "eval_runtime": 4.0548,
+      "eval_samples_per_second": 24.662,
+      "eval_steps_per_second": 3.206,
+      "step": 56
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 1.6853038769745467e-05,
+      "loss": 0.1663,
+      "step": 60
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 5.8455934816979305e-06,
+      "loss": 0.1468,
+      "step": 65
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 4.815273327803182e-07,
+      "loss": 0.1591,
+      "step": 70
+    },
+    {
+      "epoch": 3.84,
+      "eval_loss": 0.21221186220645905,
+      "eval_runtime": 4.0724,
+      "eval_samples_per_second": 24.556,
+      "eval_steps_per_second": 3.192,
+      "step": 72
+    },
+    {
+      "epoch": 3.84,
+      "step": 72,
+      "total_flos": 40669301473280.0,
+      "train_loss": 0.4049788423710399,
+      "train_runtime": 399.0811,
+      "train_samples_per_second": 3.007,
+      "train_steps_per_second": 0.18
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 72,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 4,
+  "save_steps": 500,
+  "total_flos": 40669301473280.0,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}