Training in progress, step 20000, checkpoint

checkpoint-20000/README.md

@@ -0,0 +1,21 @@
+---
+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
+### Framework versions
+
+
+- PEFT 0.6.0.dev0

checkpoint-20000/adapter_config.json

@@ -0,0 +1,21 @@
+{
+  "auto_mapping": null,
+  "base_model_name_or_path": "codellama/CodeLlama-13b-Instruct-hf",
+  "bias": "lora_only",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "lora_alpha": 8,
+  "lora_dropout": 0.1,
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 8,
+  "revision": null,
+  "target_modules": [
+    "q_proj",
+    "v_proj"
+  ],
+  "task_type": "CAUSAL_LM"
+}

checkpoint-20000/adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5774dcbb58111b34e8e08e674a78eee5eea306f46a2dac7c748d6203c2c7c23d
+size 26271757

checkpoint-20000/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c2ea6e4912364724ac11e45fa54ae6fb6a3e67643850476a404c6713db3b6320
+size 13729605

checkpoint-20000/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b24ccc08df4bf4ae5396369d483a6462b10aa025f8d6d164ba5855e71213e98
+size 14575

checkpoint-20000/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2c24b584dcf77bfb1f7b85955600c58dd77cc34b5ad90a30bfc3feaeae3f876d
+size 627

checkpoint-20000/trainer_state.json

@@ -0,0 +1,225 @@
+{
+  "best_metric": 0.020885193720459938,
+  "best_model_checkpoint": "./text2sql/13b-codellama_instruct_pt_text2sql/checkpoint-20000",
+  "epoch": 2.22796231958727,
+  "eval_steps": 2000,
+  "global_step": 20000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.1139992573338283e-07,
+      "loss": 1.0897,
+      "step": 1
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 0.00011139992573338284,
+      "loss": 0.2546,
+      "step": 1000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 0.00022279985146676567,
+      "loss": 0.0774,
+      "step": 2000
+    },
+    {
+      "epoch": 0.22,
+      "eval_loss": 0.0608125701546669,
+      "eval_runtime": 285.1249,
+      "eval_samples_per_second": 7.014,
+      "eval_steps_per_second": 1.754,
+      "step": 2000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 0.00029781686301467276,
+      "loss": 0.0652,
+      "step": 3000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 0.0002907056676227274,
+      "loss": 0.0581,
+      "step": 4000
+    },
+    {
+      "epoch": 0.45,
+      "eval_loss": 0.04370184242725372,
+      "eval_runtime": 284.9797,
+      "eval_samples_per_second": 7.018,
+      "eval_steps_per_second": 1.755,
+      "step": 4000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 0.00028359447223078195,
+      "loss": 0.0499,
+      "step": 5000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 0.0002764832768388366,
+      "loss": 0.0445,
+      "step": 6000
+    },
+    {
+      "epoch": 0.67,
+      "eval_loss": 0.036812592297792435,
+      "eval_runtime": 284.5683,
+      "eval_samples_per_second": 7.028,
+      "eval_steps_per_second": 1.757,
+      "step": 6000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 0.0002693720814468912,
+      "loss": 0.0433,
+      "step": 7000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 0.00026226088605494583,
+      "loss": 0.0378,
+      "step": 8000
+    },
+    {
+      "epoch": 0.89,
+      "eval_loss": 0.03097330778837204,
+      "eval_runtime": 354.6673,
+      "eval_samples_per_second": 5.639,
+      "eval_steps_per_second": 1.41,
+      "step": 8000
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 0.00025514969066300046,
+      "loss": 0.0364,
+      "step": 9000
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 0.00024803849527105503,
+      "loss": 0.0312,
+      "step": 10000
+    },
+    {
+      "epoch": 1.11,
+      "eval_loss": 0.02841157279908657,
+      "eval_runtime": 355.7754,
+      "eval_samples_per_second": 5.622,
+      "eval_steps_per_second": 1.405,
+      "step": 10000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 0.00024092729987910966,
+      "loss": 0.0307,
+      "step": 11000
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 0.00023381610448716428,
+      "loss": 0.0316,
+      "step": 12000
+    },
+    {
+      "epoch": 1.34,
+      "eval_loss": 0.02577413246035576,
+      "eval_runtime": 284.3906,
+      "eval_samples_per_second": 7.033,
+      "eval_steps_per_second": 1.758,
+      "step": 12000
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 0.00022670490909521888,
+      "loss": 0.0272,
+      "step": 13000
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 0.0002195937137032735,
+      "loss": 0.0279,
+      "step": 14000
+    },
+    {
+      "epoch": 1.56,
+      "eval_loss": 0.024905819445848465,
+      "eval_runtime": 285.1649,
+      "eval_samples_per_second": 7.013,
+      "eval_steps_per_second": 1.753,
+      "step": 14000
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 0.00021248251831132813,
+      "loss": 0.0285,
+      "step": 15000
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 0.00020537132291938273,
+      "loss": 0.0266,
+      "step": 16000
+    },
+    {
+      "epoch": 1.78,
+      "eval_loss": 0.023692620918154716,
+      "eval_runtime": 285.1334,
+      "eval_samples_per_second": 7.014,
+      "eval_steps_per_second": 1.754,
+      "step": 16000
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 0.00019826012752743733,
+      "loss": 0.0266,
+      "step": 17000
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 0.00019114893213549196,
+      "loss": 0.0258,
+      "step": 18000
+    },
+    {
+      "epoch": 2.01,
+      "eval_loss": 0.021677831187844276,
+      "eval_runtime": 285.282,
+      "eval_samples_per_second": 7.011,
+      "eval_steps_per_second": 1.753,
+      "step": 18000
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 0.00018403773674354658,
+      "loss": 0.0209,
+      "step": 19000
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 0.0001769265413516012,
+      "loss": 0.0211,
+      "step": 20000
+    },
+    {
+      "epoch": 2.23,
+      "eval_loss": 0.020885193720459938,
+      "eval_runtime": 284.9168,
+      "eval_samples_per_second": 7.02,
+      "eval_steps_per_second": 1.755,
+      "step": 20000
+    }
+  ],
+  "logging_steps": 1000,
+  "max_steps": 44880,
+  "num_train_epochs": 5,
+  "save_steps": 2000,
+  "total_flos": 1.3760008859838136e+19,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-20000/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a01f3a3b51470d76c911f16f658a5b2f1c2fa53e2254151c1cfd84b69dd4c6c9
+size 4283