Upload folder using huggingface_hub

README.md

@@ -0,0 +1,23 @@
+---
+library_name: peft
+---
+## Training procedure
+
+
+The following `bitsandbytes` quantization config was used during training:
+- quant_method: bitsandbytes
+- _load_in_8bit: False
+- _load_in_4bit: True
+- llm_int8_threshold: 6.0
+- llm_int8_skip_modules: None
+- llm_int8_enable_fp32_cpu_offload: False
+- llm_int8_has_fp16_weight: False
+- bnb_4bit_quant_type: nf4
+- bnb_4bit_use_double_quant: True
+- bnb_4bit_compute_dtype: bfloat16
+- load_in_4bit: True
+- load_in_8bit: False
+### Framework versions
+
+
+- PEFT 0.4.0

adapter_config.json

@@ -0,0 +1,26 @@
+{
+  "auto_mapping": null,
+  "base_model_name_or_path": "meta-llama/Llama-2-7b-hf",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "lora_alpha": 64,
+  "lora_dropout": 0.1,
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 16,
+  "revision": null,
+  "target_modules": [
+    "up_proj",
+    "gate_proj",
+    "v_proj",
+    "q_proj",
+    "o_proj",
+    "down_proj",
+    "k_proj"
+  ],
+  "task_type": "SEQ_CLS"
+}

adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:44dca49c744beb5b5154de61adad3fa409959e53d3efe0b94d4b6b4be819c8c7
+size 160102926

all_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 1.0,
+    "total_flos": 1.0505428534296576e+17,
+    "train_loss": 0.5505467119013391,
+    "train_runtime": 36020.1073,
+    "train_samples_per_second": 0.146,
+    "train_steps_per_second": 0.002
+}

runs/Mar19_18-58-37_11dfe7fa700f/events.out.tfevents.1710874719.11dfe7fa700f.34.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a4e3a4eff3089375f416dc06bf51bacc2cc57bc459126a71f61ebf43c9de370a
+size 12937

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 1.0,
+    "total_flos": 1.0505428534296576e+17,
+    "train_loss": 0.5505467119013391,
+    "train_runtime": 36020.1073,
+    "train_samples_per_second": 0.146,
+    "train_steps_per_second": 0.002
+}

trainer_state.json

@@ -0,0 +1,230 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.9969604863221885,
+  "eval_steps": 8,
+  "global_step": 82,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.1,
+      "grad_norm": 12.02409839630127,
+      "learning_rate": 1.8500000000000002e-05,
+      "loss": 1.0056,
+      "step": 8
+    },
+    {
+      "epoch": 0.1,
+      "eval_accuracy": 0.76,
+      "eval_f1_score": 0.76,
+      "eval_gmean": 0.7536682052115493,
+      "eval_loss": 0.6476712226867676,
+      "eval_precision": 0.76,
+      "eval_recall": 0.76,
+      "eval_runtime": 211.0576,
+      "eval_samples_per_second": 0.474,
+      "eval_steps_per_second": 0.062,
+      "step": 8
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 13.794071197509766,
+      "learning_rate": 1.65e-05,
+      "loss": 0.761,
+      "step": 16
+    },
+    {
+      "epoch": 0.19,
+      "eval_accuracy": 0.8,
+      "eval_f1_score": 0.8,
+      "eval_gmean": 0.7947281139113244,
+      "eval_loss": 0.5526618361473083,
+      "eval_precision": 0.8,
+      "eval_recall": 0.8,
+      "eval_runtime": 211.7038,
+      "eval_samples_per_second": 0.472,
+      "eval_steps_per_second": 0.061,
+      "step": 16
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 9.750018119812012,
+      "learning_rate": 1.45e-05,
+      "loss": 0.638,
+      "step": 24
+    },
+    {
+      "epoch": 0.29,
+      "eval_accuracy": 0.81,
+      "eval_f1_score": 0.8102915601023019,
+      "eval_gmean": 0.8066432910616553,
+      "eval_loss": 0.4845515191555023,
+      "eval_precision": 0.8107466340269278,
+      "eval_recall": 0.81,
+      "eval_runtime": 211.1147,
+      "eval_samples_per_second": 0.474,
+      "eval_steps_per_second": 0.062,
+      "step": 24
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 9.329747200012207,
+      "learning_rate": 1.25e-05,
+      "loss": 0.6367,
+      "step": 32
+    },
+    {
+      "epoch": 0.39,
+      "eval_accuracy": 0.81,
+      "eval_f1_score": 0.8102915601023019,
+      "eval_gmean": 0.8066432910616553,
+      "eval_loss": 0.4284454584121704,
+      "eval_precision": 0.8107466340269278,
+      "eval_recall": 0.81,
+      "eval_runtime": 210.7064,
+      "eval_samples_per_second": 0.475,
+      "eval_steps_per_second": 0.062,
+      "step": 32
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 9.867768287658691,
+      "learning_rate": 1.0500000000000001e-05,
+      "loss": 0.5932,
+      "step": 40
+    },
+    {
+      "epoch": 0.49,
+      "eval_accuracy": 0.82,
+      "eval_f1_score": 0.82,
+      "eval_gmean": 0.8152572201307815,
+      "eval_loss": 0.3858851194381714,
+      "eval_precision": 0.82,
+      "eval_recall": 0.82,
+      "eval_runtime": 211.0479,
+      "eval_samples_per_second": 0.474,
+      "eval_steps_per_second": 0.062,
+      "step": 40
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 10.276468276977539,
+      "learning_rate": 8.5e-06,
+      "loss": 0.4603,
+      "step": 48
+    },
+    {
+      "epoch": 0.58,
+      "eval_accuracy": 0.83,
+      "eval_f1_score": 0.8297023993409536,
+      "eval_gmean": 0.823868981405519,
+      "eval_loss": 0.3537224233150482,
+      "eval_precision": 0.8295742042166185,
+      "eval_recall": 0.83,
+      "eval_runtime": 212.1782,
+      "eval_samples_per_second": 0.471,
+      "eval_steps_per_second": 0.061,
+      "step": 48
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 8.40848445892334,
+      "learning_rate": 6.5000000000000004e-06,
+      "loss": 0.3794,
+      "step": 56
+    },
+    {
+      "epoch": 0.68,
+      "eval_accuracy": 0.86,
+      "eval_f1_score": 0.86,
+      "eval_gmean": 0.8563139396380225,
+      "eval_loss": 0.3282211720943451,
+      "eval_precision": 0.86,
+      "eval_recall": 0.86,
+      "eval_runtime": 211.0187,
+      "eval_samples_per_second": 0.474,
+      "eval_steps_per_second": 0.062,
+      "step": 56
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 8.91843032836914,
+      "learning_rate": 4.5e-06,
+      "loss": 0.4213,
+      "step": 64
+    },
+    {
+      "epoch": 0.78,
+      "eval_accuracy": 0.87,
+      "eval_f1_score": 0.8701994884910484,
+      "eval_gmean": 0.8682247898524239,
+      "eval_loss": 0.31104254722595215,
+      "eval_precision": 0.8705752753977967,
+      "eval_recall": 0.87,
+      "eval_runtime": 211.0456,
+      "eval_samples_per_second": 0.474,
+      "eval_steps_per_second": 0.062,
+      "step": 64
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 6.700900554656982,
+      "learning_rate": 2.5e-06,
+      "loss": 0.3611,
+      "step": 72
+    },
+    {
+      "epoch": 0.88,
+      "eval_accuracy": 0.88,
+      "eval_f1_score": 0.8803427172582619,
+      "eval_gmean": 0.8801313530867828,
+      "eval_loss": 0.30110642313957214,
+      "eval_precision": 0.881396103896104,
+      "eval_recall": 0.88,
+      "eval_runtime": 211.6438,
+      "eval_samples_per_second": 0.472,
+      "eval_steps_per_second": 0.061,
+      "step": 72
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 7.661783695220947,
+      "learning_rate": 5.000000000000001e-07,
+      "loss": 0.3248,
+      "step": 80
+    },
+    {
+      "epoch": 0.97,
+      "eval_accuracy": 0.88,
+      "eval_f1_score": 0.8803427172582619,
+      "eval_gmean": 0.8801313530867828,
+      "eval_loss": 0.29578596353530884,
+      "eval_precision": 0.881396103896104,
+      "eval_recall": 0.88,
+      "eval_runtime": 210.1348,
+      "eval_samples_per_second": 0.476,
+      "eval_steps_per_second": 0.062,
+      "step": 80
+    },
+    {
+      "epoch": 1.0,
+      "step": 82,
+      "total_flos": 1.0505428534296576e+17,
+      "train_loss": 0.5505467119013391,
+      "train_runtime": 36020.1073,
+      "train_samples_per_second": 0.146,
+      "train_steps_per_second": 0.002
+    }
+  ],
+  "logging_steps": 8,
+  "max_steps": 82,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 500,
+  "total_flos": 1.0505428534296576e+17,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}