Training in progress, step 400

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:809ad0e34ad8de3ae8361ae26ee722fdabe99a3bf236af9e391fff6cd2b190d7
+oid sha256:ce801567d87fbf8101c2e17b1a4088b2f2095b232559ebaf6de45abcdf3cea3e
 size 319977229

checkpoint-200/adapter_model/adapter_model/README.md

@@ -0,0 +1,20 @@
+---
+library_name: peft
+---
+## Training procedure
+
+
+The following `bitsandbytes` quantization config was used during training:
+- 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 0.4.0

checkpoint-200/adapter_model/adapter_model/adapter_config.json

@@ -0,0 +1,26 @@
+{
+  "auto_mapping": null,
+  "base_model_name_or_path": "pankajmathur/orca_mini_v3_7b",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "lora_alpha": 16.0,
+  "lora_dropout": 0.1,
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 64,
+  "revision": null,
+  "target_modules": [
+    "gate_proj",
+    "k_proj",
+    "down_proj",
+    "up_proj",
+    "o_proj",
+    "v_proj",
+    "q_proj"
+  ],
+  "task_type": "CAUSAL_LM"
+}

checkpoint-200/adapter_model/adapter_model/adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:809ad0e34ad8de3ae8361ae26ee722fdabe99a3bf236af9e391fff6cd2b190d7
+size 319977229

checkpoint-400/README.md

@@ -0,0 +1,20 @@
+---
+library_name: peft
+---
+## Training procedure
+
+
+The following `bitsandbytes` quantization config was used during training:
+- 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 0.4.0

checkpoint-400/adapter_config.json

@@ -0,0 +1,26 @@
+{
+  "auto_mapping": null,
+  "base_model_name_or_path": "pankajmathur/orca_mini_v3_7b",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "lora_alpha": 16.0,
+  "lora_dropout": 0.1,
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 64,
+  "revision": null,
+  "target_modules": [
+    "gate_proj",
+    "k_proj",
+    "down_proj",
+    "up_proj",
+    "o_proj",
+    "v_proj",
+    "q_proj"
+  ],
+  "task_type": "CAUSAL_LM"
+}

checkpoint-400/adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ce801567d87fbf8101c2e17b1a4088b2f2095b232559ebaf6de45abcdf3cea3e
+size 319977229

checkpoint-400/added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "[PAD]": 32000
+}

checkpoint-400/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ff1feb73d38cf64e9de27150be355c54bb57157ad521c045d077406a5e8aac04
+size 1279539973

checkpoint-400/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5a9f7b8f8d2fb4e1d69418c5b8af45a31b4a2aab5eba80a71c92ac9ed1dd5fe0
+size 14511

checkpoint-400/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:13276f15dd2b6acc19b970176aa2db4ac9b58241843e72c89b50e3094e903b19
+size 627

checkpoint-400/special_tokens_map.json

@@ -0,0 +1,6 @@
+{
+  "bos_token": "<s>",
+  "eos_token": "</s>",
+  "pad_token": "[PAD]",
+  "unk_token": "<unk>"
+}

checkpoint-400/tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
+size 499723

checkpoint-400/tokenizer_config.json

@@ -0,0 +1,35 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "bos_token": {
+    "__type": "AddedToken",
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "clean_up_tokenization_spaces": false,
+  "eos_token": {
+    "__type": "AddedToken",
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  },
+  "legacy": true,
+  "model_max_length": 4096,
+  "pad_token": null,
+  "padding_side": "right",
+  "sp_model_kwargs": {},
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": {
+    "__type": "AddedToken",
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": true,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-400/trainer_state.json

@@ -0,0 +1,398 @@
+{
+  "best_metric": 0.761216402053833,
+  "best_model_checkpoint": "experts/expert-4/checkpoint-400",
+  "epoch": 0.0447152199429881,
+  "global_step": 400,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0002,
+      "loss": 0.831,
+      "step": 10
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0002,
+      "loss": 0.9543,
+      "step": 20
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0002,
+      "loss": 0.8393,
+      "step": 30
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 0.0002,
+      "loss": 0.8524,
+      "step": 40
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0002,
+      "loss": 0.8036,
+      "step": 50
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0002,
+      "loss": 0.9743,
+      "step": 60
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0002,
+      "loss": 0.7541,
+      "step": 70
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0002,
+      "loss": 0.6946,
+      "step": 80
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0002,
+      "loss": 0.8065,
+      "step": 90
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0002,
+      "loss": 0.9007,
+      "step": 100
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0002,
+      "loss": 0.8615,
+      "step": 110
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0002,
+      "loss": 0.7233,
+      "step": 120
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 0.0002,
+      "loss": 0.8257,
+      "step": 130
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.0002,
+      "loss": 0.8099,
+      "step": 140
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.0002,
+      "loss": 0.7405,
+      "step": 150
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.0002,
+      "loss": 0.8419,
+      "step": 160
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.0002,
+      "loss": 0.7397,
+      "step": 170
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.0002,
+      "loss": 0.7997,
+      "step": 180
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.0002,
+      "loss": 0.8696,
+      "step": 190
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.0002,
+      "loss": 0.8595,
+      "step": 200
+    },
+    {
+      "epoch": 0.02,
+      "eval_loss": 0.7751898169517517,
+      "eval_runtime": 133.9181,
+      "eval_samples_per_second": 7.467,
+      "eval_steps_per_second": 3.734,
+      "step": 200
+    },
+    {
+      "epoch": 0.02,
+      "mmlu_eval_accuracy": 0.48961042276821043,
+      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
+      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
+      "mmlu_eval_accuracy_astronomy": 0.4375,
+      "mmlu_eval_accuracy_business_ethics": 0.45454545454545453,
+      "mmlu_eval_accuracy_clinical_knowledge": 0.5517241379310345,
+      "mmlu_eval_accuracy_college_biology": 0.3125,
+      "mmlu_eval_accuracy_college_chemistry": 0.25,
+      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
+      "mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
+      "mmlu_eval_accuracy_college_medicine": 0.45454545454545453,
+      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
+      "mmlu_eval_accuracy_computer_security": 0.45454545454545453,
+      "mmlu_eval_accuracy_conceptual_physics": 0.38461538461538464,
+      "mmlu_eval_accuracy_econometrics": 0.25,
+      "mmlu_eval_accuracy_electrical_engineering": 0.25,
+      "mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536,
+      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
+      "mmlu_eval_accuracy_global_facts": 0.3,
+      "mmlu_eval_accuracy_high_school_biology": 0.375,
+      "mmlu_eval_accuracy_high_school_chemistry": 0.3181818181818182,
+      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
+      "mmlu_eval_accuracy_high_school_european_history": 0.6111111111111112,
+      "mmlu_eval_accuracy_high_school_geography": 0.9090909090909091,
+      "mmlu_eval_accuracy_high_school_government_and_politics": 0.5714285714285714,
+      "mmlu_eval_accuracy_high_school_macroeconomics": 0.46511627906976744,
+      "mmlu_eval_accuracy_high_school_mathematics": 0.2413793103448276,
+      "mmlu_eval_accuracy_high_school_microeconomics": 0.5,
+      "mmlu_eval_accuracy_high_school_physics": 0.29411764705882354,
+      "mmlu_eval_accuracy_high_school_psychology": 0.85,
+      "mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
+      "mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
+      "mmlu_eval_accuracy_high_school_world_history": 0.7307692307692307,
+      "mmlu_eval_accuracy_human_aging": 0.6521739130434783,
+      "mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
+      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
+      "mmlu_eval_accuracy_jurisprudence": 0.45454545454545453,
+      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
+      "mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
+      "mmlu_eval_accuracy_management": 0.6363636363636364,
+      "mmlu_eval_accuracy_marketing": 0.8,
+      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
+      "mmlu_eval_accuracy_miscellaneous": 0.6744186046511628,
+      "mmlu_eval_accuracy_moral_disputes": 0.3684210526315789,
+      "mmlu_eval_accuracy_moral_scenarios": 0.24,
+      "mmlu_eval_accuracy_nutrition": 0.6363636363636364,
+      "mmlu_eval_accuracy_philosophy": 0.5,
+      "mmlu_eval_accuracy_prehistory": 0.5428571428571428,
+      "mmlu_eval_accuracy_professional_accounting": 0.2903225806451613,
+      "mmlu_eval_accuracy_professional_law": 0.3235294117647059,
+      "mmlu_eval_accuracy_professional_medicine": 0.5483870967741935,
+      "mmlu_eval_accuracy_professional_psychology": 0.4927536231884058,
+      "mmlu_eval_accuracy_public_relations": 0.5833333333333334,
+      "mmlu_eval_accuracy_security_studies": 0.48148148148148145,
+      "mmlu_eval_accuracy_sociology": 0.7272727272727273,
+      "mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
+      "mmlu_eval_accuracy_virology": 0.4444444444444444,
+      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
+      "mmlu_loss": 1.4005114495287365,
+      "step": 200
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.0002,
+      "loss": 0.8483,
+      "step": 210
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 0.0002,
+      "loss": 0.8109,
+      "step": 220
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.0002,
+      "loss": 0.7124,
+      "step": 230
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.0002,
+      "loss": 0.8426,
+      "step": 240
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.0002,
+      "loss": 0.7161,
+      "step": 250
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.0002,
+      "loss": 0.9172,
+      "step": 260
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.0002,
+      "loss": 0.8112,
+      "step": 270
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.0002,
+      "loss": 0.7547,
+      "step": 280
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.0002,
+      "loss": 0.7454,
+      "step": 290
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.0002,
+      "loss": 0.7149,
+      "step": 300
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 0.0002,
+      "loss": 0.8117,
+      "step": 310
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.0002,
+      "loss": 0.7332,
+      "step": 320
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.0002,
+      "loss": 0.7839,
+      "step": 330
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.0002,
+      "loss": 0.8199,
+      "step": 340
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.0002,
+      "loss": 0.799,
+      "step": 350
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.0002,
+      "loss": 0.7754,
+      "step": 360
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.0002,
+      "loss": 0.7456,
+      "step": 370
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.0002,
+      "loss": 0.8303,
+      "step": 380
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.0002,
+      "loss": 0.7955,
+      "step": 390
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 0.0002,
+      "loss": 0.7828,
+      "step": 400
+    },
+    {
+      "epoch": 0.04,
+      "eval_loss": 0.761216402053833,
+      "eval_runtime": 133.828,
+      "eval_samples_per_second": 7.472,
+      "eval_steps_per_second": 3.736,
+      "step": 400
+    },
+    {
+      "epoch": 0.04,
+      "mmlu_eval_accuracy": 0.49707868719131965,
+      "mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
+      "mmlu_eval_accuracy_anatomy": 0.5714285714285714,
+      "mmlu_eval_accuracy_astronomy": 0.5,
+      "mmlu_eval_accuracy_business_ethics": 0.45454545454545453,
+      "mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
+      "mmlu_eval_accuracy_college_biology": 0.25,
+      "mmlu_eval_accuracy_college_chemistry": 0.375,
+      "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
+      "mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
+      "mmlu_eval_accuracy_college_medicine": 0.45454545454545453,
+      "mmlu_eval_accuracy_college_physics": 0.36363636363636365,
+      "mmlu_eval_accuracy_computer_security": 0.5454545454545454,
+      "mmlu_eval_accuracy_conceptual_physics": 0.46153846153846156,
+      "mmlu_eval_accuracy_econometrics": 0.25,
+      "mmlu_eval_accuracy_electrical_engineering": 0.25,
+      "mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536,
+      "mmlu_eval_accuracy_formal_logic": 0.35714285714285715,
+      "mmlu_eval_accuracy_global_facts": 0.4,
+      "mmlu_eval_accuracy_high_school_biology": 0.40625,
+      "mmlu_eval_accuracy_high_school_chemistry": 0.2727272727272727,
+      "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
+      "mmlu_eval_accuracy_high_school_european_history": 0.6666666666666666,
+      "mmlu_eval_accuracy_high_school_geography": 0.9090909090909091,
+      "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
+      "mmlu_eval_accuracy_high_school_macroeconomics": 0.46511627906976744,
+      "mmlu_eval_accuracy_high_school_mathematics": 0.1724137931034483,
+      "mmlu_eval_accuracy_high_school_microeconomics": 0.5384615384615384,
+      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
+      "mmlu_eval_accuracy_high_school_psychology": 0.85,
+      "mmlu_eval_accuracy_high_school_statistics": 0.43478260869565216,
+      "mmlu_eval_accuracy_high_school_us_history": 0.6363636363636364,
+      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
+      "mmlu_eval_accuracy_human_aging": 0.6521739130434783,
+      "mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
+      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
+      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
+      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
+      "mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
+      "mmlu_eval_accuracy_management": 0.6363636363636364,
+      "mmlu_eval_accuracy_marketing": 0.8,
+      "mmlu_eval_accuracy_medical_genetics": 0.8181818181818182,
+      "mmlu_eval_accuracy_miscellaneous": 0.6511627906976745,
+      "mmlu_eval_accuracy_moral_disputes": 0.42105263157894735,
+      "mmlu_eval_accuracy_moral_scenarios": 0.23,
+      "mmlu_eval_accuracy_nutrition": 0.7575757575757576,
+      "mmlu_eval_accuracy_philosophy": 0.5,
+      "mmlu_eval_accuracy_prehistory": 0.4857142857142857,
+      "mmlu_eval_accuracy_professional_accounting": 0.25806451612903225,
+      "mmlu_eval_accuracy_professional_law": 0.35294117647058826,
+      "mmlu_eval_accuracy_professional_medicine": 0.5483870967741935,
+      "mmlu_eval_accuracy_professional_psychology": 0.4782608695652174,
+      "mmlu_eval_accuracy_public_relations": 0.6666666666666666,
+      "mmlu_eval_accuracy_security_studies": 0.48148148148148145,
+      "mmlu_eval_accuracy_sociology": 0.7272727272727273,
+      "mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
+      "mmlu_eval_accuracy_virology": 0.4444444444444444,
+      "mmlu_eval_accuracy_world_religions": 0.6842105263157895,
+      "mmlu_loss": 1.2046297034224394,
+      "step": 400
+    }
+  ],
+  "max_steps": 10000,
+  "num_train_epochs": 2,
+  "total_flos": 4.044337437592781e+16,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-400/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7005548a7ac40ea8858de7051ca7cff78e2d6f2c2717517ade58fcdd6ce38912
+size 5819