upload loras

llama2_13b_peft/logical_deduction/README.md

@@ -0,0 +1,73 @@
+---
+license: other
+library_name: peft
+tags:
+- llama-factory
+- lora
+- generated_from_trainer
+base_model: /data1/model/llama2/meta-llama/Llama2-13b
+model-index:
+- name: logical_deduction_no_sys
+  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. -->
+
+# logical_deduction_no_sys
+
+This model is a fine-tuned version of [/data1/model/llama2/meta-llama/Llama2-13b](https://huggingface.co//data1/model/llama2/meta-llama/Llama2-13b) on the logical_deduction_no_sys dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.1318
+
+## 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: 0.0001
+- train_batch_size: 8
+- eval_batch_size: 8
+- seed: 42
+- distributed_type: multi-GPU
+- num_devices: 3
+- total_train_batch_size: 24
+- total_eval_batch_size: 24
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: cosine
+- lr_scheduler_warmup_steps: 20
+- num_epochs: 10.0
+
+### Training results
+
+| Training Loss | Epoch | Step | Validation Loss |
+|:-------------:|:-----:|:----:|:---------------:|
+| 0.121         | 1.16  | 50   | 0.1113          |
+| 0.0574        | 2.33  | 100  | 0.0727          |
+| 0.0405        | 3.49  | 150  | 0.0608          |
+| 0.0319        | 4.65  | 200  | 0.0899          |
+| 0.0236        | 5.81  | 250  | 0.0640          |
+| 0.0152        | 6.98  | 300  | 0.0996          |
+| 0.0053        | 8.14  | 350  | 0.1150          |
+| 0.0021        | 9.3   | 400  | 0.1312          |
+
+
+### Framework versions
+
+- PEFT 0.9.0
+- Transformers 4.38.2
+- Pytorch 2.2.1
+- Datasets 2.18.0
+- Tokenizers 0.15.2

llama2_13b_peft/logical_deduction/adapter_config.json

@@ -0,0 +1,33 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data1/model/llama2/meta-llama/Llama2-13b",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_dropout": 0.0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "gate_proj",
+    "k_proj",
+    "q_proj",
+    "down_proj",
+    "up_proj",
+    "o_proj",
+    "v_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

llama2_13b_peft/logical_deduction/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:52d94d189237fab0fa3b831e5e3d81d2f34e9070ecf232ca2d1a17d2b0faee80
+size 125248064

llama2_13b_peft/logical_deduction/all_results.json

@@ -0,0 +1,11 @@
+{
+    "epoch": 10.0,
+    "eval_loss": 0.1317996233701706,
+    "eval_runtime": 2.219,
+    "eval_samples_per_second": 81.118,
+    "eval_steps_per_second": 3.605,
+    "train_loss": 0.0926927653899373,
+    "train_runtime": 420.9269,
+    "train_samples_per_second": 24.232,
+    "train_steps_per_second": 1.022
+}

llama2_13b_peft/logical_deduction/eval_results.json

@@ -0,0 +1,7 @@
+{
+    "epoch": 10.0,
+    "eval_loss": 0.1317996233701706,
+    "eval_runtime": 2.219,
+    "eval_samples_per_second": 81.118,
+    "eval_steps_per_second": 3.605
+}

llama2_13b_peft/logical_deduction/special_tokens_map.json

@@ -0,0 +1,24 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "</s>",
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

llama2_13b_peft/logical_deduction/tokenizer.model

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

llama2_13b_peft/logical_deduction/tokenizer_config.json

@@ -0,0 +1,45 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "add_prefix_space": true,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "chat_template": "{% if messages[0]['role'] == 'system' %}{% set system_message = messages[0]['content'] %}{% endif %}{% if system_message is defined %}{{ system_message }}{% endif %}{% for message in messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ content }}{% elif message['role'] == 'assistant' %}{{ content + '\\n' }}{% endif %}{% endfor %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "</s>",
+  "legacy": true,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "</s>",
+  "padding_side": "right",
+  "sp_model_kwargs": {},
+  "spaces_between_special_tokens": false,
+  "split_special_tokens": false,
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": false
+}

llama2_13b_peft/logical_deduction/train_results.json

@@ -0,0 +1,7 @@
+{
+    "epoch": 10.0,
+    "train_loss": 0.0926927653899373,
+    "train_runtime": 420.9269,
+    "train_samples_per_second": 24.232,
+    "train_steps_per_second": 1.022
+}

llama2_13b_peft/logical_deduction/trainer_log.jsonl

@@ -0,0 +1,53 @@
+{"current_steps": 10, "total_steps": 430, "loss": 1.8041, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 5e-05, "epoch": 0.23, "percentage": 2.33, "elapsed_time": "0:00:10", "remaining_time": "0:07:22"}
+{"current_steps": 20, "total_steps": 430, "loss": 0.5814, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 0.0001, "epoch": 0.47, "percentage": 4.65, "elapsed_time": "0:00:19", "remaining_time": "0:06:42"}
+{"current_steps": 30, "total_steps": 430, "loss": 0.1611, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 9.985329005918702e-05, "epoch": 0.7, "percentage": 6.98, "elapsed_time": "0:00:28", "remaining_time": "0:06:21"}
+{"current_steps": 40, "total_steps": 430, "loss": 0.1289, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 9.941402118901744e-05, "epoch": 0.93, "percentage": 9.3, "elapsed_time": "0:00:37", "remaining_time": "0:06:09"}
+{"current_steps": 50, "total_steps": 430, "loss": 0.121, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 9.868477119388896e-05, "epoch": 1.16, "percentage": 11.63, "elapsed_time": "0:00:47", "remaining_time": "0:05:58"}
+{"current_steps": 50, "total_steps": 430, "loss": null, "eval_loss": 0.11131427437067032, "predict_loss": null, "reward": null, "learning_rate": null, "epoch": 1.16, "percentage": 11.63, "elapsed_time": "0:00:47", "remaining_time": "0:05:58"}
+{"current_steps": 60, "total_steps": 430, "loss": 0.1046, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 9.766981960274653e-05, "epoch": 1.4, "percentage": 13.95, "elapsed_time": "0:00:58", "remaining_time": "0:06:01"}
+{"current_steps": 70, "total_steps": 430, "loss": 0.0922, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 9.637512255510475e-05, "epoch": 1.63, "percentage": 16.28, "elapsed_time": "0:01:07", "remaining_time": "0:05:49"}
+{"current_steps": 80, "total_steps": 430, "loss": 0.0957, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 9.480827784805278e-05, "epoch": 1.86, "percentage": 18.6, "elapsed_time": "0:01:16", "remaining_time": "0:05:36"}
+{"current_steps": 90, "total_steps": 430, "loss": 0.0751, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 9.297848034936006e-05, "epoch": 2.09, "percentage": 20.93, "elapsed_time": "0:01:26", "remaining_time": "0:05:24"}
+{"current_steps": 100, "total_steps": 430, "loss": 0.0574, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 9.089646803833589e-05, "epoch": 2.33, "percentage": 23.26, "elapsed_time": "0:01:35", "remaining_time": "0:05:14"}
+{"current_steps": 100, "total_steps": 430, "loss": null, "eval_loss": 0.07272274047136307, "predict_loss": null, "reward": null, "learning_rate": null, "epoch": 2.33, "percentage": 23.26, "elapsed_time": "0:01:35", "remaining_time": "0:05:14"}
+{"current_steps": 110, "total_steps": 430, "loss": 0.0602, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 8.857445899109715e-05, "epoch": 2.56, "percentage": 25.58, "elapsed_time": "0:01:46", "remaining_time": "0:05:10"}
+{"current_steps": 120, "total_steps": 430, "loss": 0.069, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 8.602607968003935e-05, "epoch": 2.79, "percentage": 27.91, "elapsed_time": "0:01:56", "remaining_time": "0:05:00"}
+{"current_steps": 130, "total_steps": 430, "loss": 0.0711, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 8.326628500827826e-05, "epoch": 3.02, "percentage": 30.23, "elapsed_time": "0:02:05", "remaining_time": "0:04:49"}
+{"current_steps": 140, "total_steps": 430, "loss": 0.0415, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 8.03112705483319e-05, "epoch": 3.26, "percentage": 32.56, "elapsed_time": "0:02:14", "remaining_time": "0:04:38"}
+{"current_steps": 150, "total_steps": 430, "loss": 0.0405, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 7.717837750006106e-05, "epoch": 3.49, "percentage": 34.88, "elapsed_time": "0:02:23", "remaining_time": "0:04:28"}
+{"current_steps": 150, "total_steps": 430, "loss": null, "eval_loss": 0.060838837176561356, "predict_loss": null, "reward": null, "learning_rate": null, "epoch": 3.49, "percentage": 34.88, "elapsed_time": "0:02:23", "remaining_time": "0:04:28"}
+{"current_steps": 160, "total_steps": 430, "loss": 0.0453, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 7.388599092561315e-05, "epoch": 3.72, "percentage": 37.21, "elapsed_time": "0:02:35", "remaining_time": "0:04:22"}
+{"current_steps": 170, "total_steps": 430, "loss": 0.0536, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 7.045343185856701e-05, "epoch": 3.95, "percentage": 39.53, "elapsed_time": "0:02:44", "remaining_time": "0:04:11"}
+{"current_steps": 180, "total_steps": 430, "loss": 0.0278, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 6.690084392042513e-05, "epoch": 4.19, "percentage": 41.86, "elapsed_time": "0:02:53", "remaining_time": "0:04:01"}
+{"current_steps": 190, "total_steps": 430, "loss": 0.034, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 6.32490751098331e-05, "epoch": 4.42, "percentage": 44.19, "elapsed_time": "0:03:03", "remaining_time": "0:03:51"}
+{"current_steps": 200, "total_steps": 430, "loss": 0.0319, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 5.951955545823342e-05, "epoch": 4.65, "percentage": 46.51, "elapsed_time": "0:03:12", "remaining_time": "0:03:41"}
+{"current_steps": 200, "total_steps": 430, "loss": null, "eval_loss": 0.08988692611455917, "predict_loss": null, "reward": null, "learning_rate": null, "epoch": 4.65, "percentage": 46.51, "elapsed_time": "0:03:12", "remaining_time": "0:03:41"}
+{"current_steps": 210, "total_steps": 430, "loss": 0.0376, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 5.573417126992003e-05, "epoch": 4.88, "percentage": 48.84, "elapsed_time": "0:03:23", "remaining_time": "0:03:33"}
+{"current_steps": 220, "total_steps": 430, "loss": 0.0308, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 5.191513668450178e-05, "epoch": 5.12, "percentage": 51.16, "elapsed_time": "0:03:32", "remaining_time": "0:03:23"}
+{"current_steps": 230, "total_steps": 430, "loss": 0.0284, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.8084863315498234e-05, "epoch": 5.35, "percentage": 53.49, "elapsed_time": "0:03:42", "remaining_time": "0:03:13"}
+{"current_steps": 240, "total_steps": 430, "loss": 0.0227, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.4265828730079987e-05, "epoch": 5.58, "percentage": 55.81, "elapsed_time": "0:03:51", "remaining_time": "0:03:03"}
+{"current_steps": 250, "total_steps": 430, "loss": 0.0236, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 4.0480444541766576e-05, "epoch": 5.81, "percentage": 58.14, "elapsed_time": "0:04:00", "remaining_time": "0:02:53"}
+{"current_steps": 250, "total_steps": 430, "loss": null, "eval_loss": 0.06400923430919647, "predict_loss": null, "reward": null, "learning_rate": null, "epoch": 5.81, "percentage": 58.14, "elapsed_time": "0:04:00", "remaining_time": "0:02:53"}
+{"current_steps": 260, "total_steps": 430, "loss": 0.0255, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 3.675092489016693e-05, "epoch": 6.05, "percentage": 60.47, "elapsed_time": "0:04:12", "remaining_time": "0:02:44"}
+{"current_steps": 270, "total_steps": 430, "loss": 0.0177, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 3.309915607957487e-05, "epoch": 6.28, "percentage": 62.79, "elapsed_time": "0:04:21", "remaining_time": "0:02:34"}
+{"current_steps": 280, "total_steps": 430, "loss": 0.0118, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 2.9546568141433006e-05, "epoch": 6.51, "percentage": 65.12, "elapsed_time": "0:04:30", "remaining_time": "0:02:24"}
+{"current_steps": 290, "total_steps": 430, "loss": 0.0128, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 2.6114009074386846e-05, "epoch": 6.74, "percentage": 67.44, "elapsed_time": "0:04:39", "remaining_time": "0:02:15"}
+{"current_steps": 300, "total_steps": 430, "loss": 0.0152, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 2.282162249993895e-05, "epoch": 6.98, "percentage": 69.77, "elapsed_time": "0:04:48", "remaining_time": "0:02:05"}
+{"current_steps": 300, "total_steps": 430, "loss": null, "eval_loss": 0.09959390014410019, "predict_loss": null, "reward": null, "learning_rate": null, "epoch": 6.98, "percentage": 69.77, "elapsed_time": "0:04:48", "remaining_time": "0:02:05"}
+{"current_steps": 310, "total_steps": 430, "loss": 0.007, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.9688729451668114e-05, "epoch": 7.21, "percentage": 72.09, "elapsed_time": "0:05:00", "remaining_time": "0:01:56"}
+{"current_steps": 320, "total_steps": 430, "loss": 0.0107, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.673371499172174e-05, "epoch": 7.44, "percentage": 74.42, "elapsed_time": "0:05:09", "remaining_time": "0:01:46"}
+{"current_steps": 330, "total_steps": 430, "loss": 0.0055, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.3973920319960655e-05, "epoch": 7.67, "percentage": 76.74, "elapsed_time": "0:05:18", "remaining_time": "0:01:36"}
+{"current_steps": 340, "total_steps": 430, "loss": 0.0116, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.1425541008902851e-05, "epoch": 7.91, "percentage": 79.07, "elapsed_time": "0:05:28", "remaining_time": "0:01:26"}
+{"current_steps": 350, "total_steps": 430, "loss": 0.0053, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 9.103531961664118e-06, "epoch": 8.14, "percentage": 81.4, "elapsed_time": "0:05:37", "remaining_time": "0:01:17"}
+{"current_steps": 350, "total_steps": 430, "loss": null, "eval_loss": 0.11503978818655014, "predict_loss": null, "reward": null, "learning_rate": null, "epoch": 8.14, "percentage": 81.4, "elapsed_time": "0:05:37", "remaining_time": "0:01:17"}
+{"current_steps": 360, "total_steps": 430, "loss": 0.0037, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 7.0215196506399515e-06, "epoch": 8.37, "percentage": 83.72, "elapsed_time": "0:05:48", "remaining_time": "0:01:07"}
+{"current_steps": 370, "total_steps": 430, "loss": 0.0034, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 5.191722151947226e-06, "epoch": 8.6, "percentage": 86.05, "elapsed_time": "0:05:57", "remaining_time": "0:00:58"}
+{"current_steps": 380, "total_steps": 430, "loss": 0.004, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 3.6248774448952695e-06, "epoch": 8.84, "percentage": 88.37, "elapsed_time": "0:06:07", "remaining_time": "0:00:48"}
+{"current_steps": 390, "total_steps": 430, "loss": 0.0032, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 2.330180397253473e-06, "epoch": 9.07, "percentage": 90.7, "elapsed_time": "0:06:16", "remaining_time": "0:00:38"}
+{"current_steps": 400, "total_steps": 430, "loss": 0.0021, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.3152288061110518e-06, "epoch": 9.3, "percentage": 93.02, "elapsed_time": "0:06:25", "remaining_time": "0:00:28"}
+{"current_steps": 400, "total_steps": 430, "loss": null, "eval_loss": 0.13117723166942596, "predict_loss": null, "reward": null, "learning_rate": null, "epoch": 9.3, "percentage": 93.02, "elapsed_time": "0:06:25", "remaining_time": "0:00:28"}
+{"current_steps": 410, "total_steps": 430, "loss": 0.0026, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 5.859788109825793e-07, "epoch": 9.53, "percentage": 95.35, "elapsed_time": "0:06:36", "remaining_time": "0:00:19"}
+{"current_steps": 420, "total_steps": 430, "loss": 0.0027, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 1.4670994081297795e-07, "epoch": 9.77, "percentage": 97.67, "elapsed_time": "0:06:46", "remaining_time": "0:00:09"}
+{"current_steps": 430, "total_steps": 430, "loss": 0.0016, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": 0.0, "epoch": 10.0, "percentage": 100.0, "elapsed_time": "0:06:55", "remaining_time": "0:00:00"}
+{"current_steps": 430, "total_steps": 430, "loss": null, "eval_loss": null, "predict_loss": null, "reward": null, "learning_rate": null, "epoch": 10.0, "percentage": 100.0, "elapsed_time": "0:06:55", "remaining_time": "0:00:00"}
+{"current_steps": 8, "total_steps": 8, "loss": null, "eval_loss": 0.1317996233701706, "predict_loss": null, "reward": null, "learning_rate": null, "epoch": 10.0, "percentage": 100.0, "elapsed_time": "0:06:57", "remaining_time": "0:00:00"}

llama2_13b_peft/logical_deduction/trainer_state.json

@@ -0,0 +1,395 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 10.0,
+  "eval_steps": 50,
+  "global_step": 430,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.23,
+      "grad_norm": 2.765148401260376,
+      "learning_rate": 5e-05,
+      "loss": 1.8041,
+      "step": 10
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 0.46828070282936096,
+      "learning_rate": 0.0001,
+      "loss": 0.5814,
+      "step": 20
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.7514229416847229,
+      "learning_rate": 9.985329005918702e-05,
+      "loss": 0.1611,
+      "step": 30
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.5662431120872498,
+      "learning_rate": 9.941402118901744e-05,
+      "loss": 0.1289,
+      "step": 40
+    },
+    {
+      "epoch": 1.16,
+      "grad_norm": 0.3684688210487366,
+      "learning_rate": 9.868477119388896e-05,
+      "loss": 0.121,
+      "step": 50
+    },
+    {
+      "epoch": 1.16,
+      "eval_loss": 0.11131427437067032,
+      "eval_runtime": 2.2112,
+      "eval_samples_per_second": 81.405,
+      "eval_steps_per_second": 3.618,
+      "step": 50
+    },
+    {
+      "epoch": 1.4,
+      "grad_norm": 0.2601034939289093,
+      "learning_rate": 9.766981960274653e-05,
+      "loss": 0.1046,
+      "step": 60
+    },
+    {
+      "epoch": 1.63,
+      "grad_norm": 0.7295863628387451,
+      "learning_rate": 9.637512255510475e-05,
+      "loss": 0.0922,
+      "step": 70
+    },
+    {
+      "epoch": 1.86,
+      "grad_norm": 0.3247099220752716,
+      "learning_rate": 9.480827784805278e-05,
+      "loss": 0.0957,
+      "step": 80
+    },
+    {
+      "epoch": 2.09,
+      "grad_norm": 0.4998391568660736,
+      "learning_rate": 9.297848034936006e-05,
+      "loss": 0.0751,
+      "step": 90
+    },
+    {
+      "epoch": 2.33,
+      "grad_norm": 0.4482787549495697,
+      "learning_rate": 9.089646803833589e-05,
+      "loss": 0.0574,
+      "step": 100
+    },
+    {
+      "epoch": 2.33,
+      "eval_loss": 0.07272274047136307,
+      "eval_runtime": 2.223,
+      "eval_samples_per_second": 80.972,
+      "eval_steps_per_second": 3.599,
+      "step": 100
+    },
+    {
+      "epoch": 2.56,
+      "grad_norm": 0.4341526925563812,
+      "learning_rate": 8.857445899109715e-05,
+      "loss": 0.0602,
+      "step": 110
+    },
+    {
+      "epoch": 2.79,
+      "grad_norm": 0.33612173795700073,
+      "learning_rate": 8.602607968003935e-05,
+      "loss": 0.069,
+      "step": 120
+    },
+    {
+      "epoch": 3.02,
+      "grad_norm": 0.21825428307056427,
+      "learning_rate": 8.326628500827826e-05,
+      "loss": 0.0711,
+      "step": 130
+    },
+    {
+      "epoch": 3.26,
+      "grad_norm": 0.5240222811698914,
+      "learning_rate": 8.03112705483319e-05,
+      "loss": 0.0415,
+      "step": 140
+    },
+    {
+      "epoch": 3.49,
+      "grad_norm": 0.7221460342407227,
+      "learning_rate": 7.717837750006106e-05,
+      "loss": 0.0405,
+      "step": 150
+    },
+    {
+      "epoch": 3.49,
+      "eval_loss": 0.060838837176561356,
+      "eval_runtime": 2.2271,
+      "eval_samples_per_second": 80.821,
+      "eval_steps_per_second": 3.592,
+      "step": 150
+    },
+    {
+      "epoch": 3.72,
+      "grad_norm": 0.26117128133773804,
+      "learning_rate": 7.388599092561315e-05,
+      "loss": 0.0453,
+      "step": 160
+    },
+    {
+      "epoch": 3.95,
+      "grad_norm": 0.41725072264671326,
+      "learning_rate": 7.045343185856701e-05,
+      "loss": 0.0536,
+      "step": 170
+    },
+    {
+      "epoch": 4.19,
+      "grad_norm": 0.3449525833129883,
+      "learning_rate": 6.690084392042513e-05,
+      "loss": 0.0278,
+      "step": 180
+    },
+    {
+      "epoch": 4.42,
+      "grad_norm": 0.9864999055862427,
+      "learning_rate": 6.32490751098331e-05,
+      "loss": 0.034,
+      "step": 190
+    },
+    {
+      "epoch": 4.65,
+      "grad_norm": 0.25210413336753845,
+      "learning_rate": 5.951955545823342e-05,
+      "loss": 0.0319,
+      "step": 200
+    },
+    {
+      "epoch": 4.65,
+      "eval_loss": 0.08988692611455917,
+      "eval_runtime": 2.2211,
+      "eval_samples_per_second": 81.039,
+      "eval_steps_per_second": 3.602,
+      "step": 200
+    },
+    {
+      "epoch": 4.88,
+      "grad_norm": 0.32543689012527466,
+      "learning_rate": 5.573417126992003e-05,
+      "loss": 0.0376,
+      "step": 210
+    },
+    {
+      "epoch": 5.12,
+      "grad_norm": 0.2424646019935608,
+      "learning_rate": 5.191513668450178e-05,
+      "loss": 0.0308,
+      "step": 220
+    },
+    {
+      "epoch": 5.35,
+      "grad_norm": 0.34920650720596313,
+      "learning_rate": 4.8084863315498234e-05,
+      "loss": 0.0284,
+      "step": 230
+    },
+    {
+      "epoch": 5.58,
+      "grad_norm": 0.4021368622779846,
+      "learning_rate": 4.4265828730079987e-05,
+      "loss": 0.0227,
+      "step": 240
+    },
+    {
+      "epoch": 5.81,
+      "grad_norm": 0.8135946989059448,
+      "learning_rate": 4.0480444541766576e-05,
+      "loss": 0.0236,
+      "step": 250
+    },
+    {
+      "epoch": 5.81,
+      "eval_loss": 0.06400923430919647,
+      "eval_runtime": 2.2241,
+      "eval_samples_per_second": 80.932,
+      "eval_steps_per_second": 3.597,
+      "step": 250
+    },
+    {
+      "epoch": 6.05,
+      "grad_norm": 0.37095046043395996,
+      "learning_rate": 3.675092489016693e-05,
+      "loss": 0.0255,
+      "step": 260
+    },
+    {
+      "epoch": 6.28,
+      "grad_norm": 0.18603980541229248,
+      "learning_rate": 3.309915607957487e-05,
+      "loss": 0.0177,
+      "step": 270
+    },
+    {
+      "epoch": 6.51,
+      "grad_norm": 0.4936978220939636,
+      "learning_rate": 2.9546568141433006e-05,
+      "loss": 0.0118,
+      "step": 280
+    },
+    {
+      "epoch": 6.74,
+      "grad_norm": 0.4769618809223175,
+      "learning_rate": 2.6114009074386846e-05,
+      "loss": 0.0128,
+      "step": 290
+    },
+    {
+      "epoch": 6.98,
+      "grad_norm": 0.4464145004749298,
+      "learning_rate": 2.282162249993895e-05,
+      "loss": 0.0152,
+      "step": 300
+    },
+    {
+      "epoch": 6.98,
+      "eval_loss": 0.09959390014410019,
+      "eval_runtime": 2.2242,
+      "eval_samples_per_second": 80.926,
+      "eval_steps_per_second": 3.597,
+      "step": 300
+    },
+    {
+      "epoch": 7.21,
+      "grad_norm": 0.12924155592918396,
+      "learning_rate": 1.9688729451668114e-05,
+      "loss": 0.007,
+      "step": 310
+    },
+    {
+      "epoch": 7.44,
+      "grad_norm": 0.4371550381183624,
+      "learning_rate": 1.673371499172174e-05,
+      "loss": 0.0107,
+      "step": 320
+    },
+    {
+      "epoch": 7.67,
+      "grad_norm": 0.23555269837379456,
+      "learning_rate": 1.3973920319960655e-05,
+      "loss": 0.0055,
+      "step": 330
+    },
+    {
+      "epoch": 7.91,
+      "grad_norm": 0.517039954662323,
+      "learning_rate": 1.1425541008902851e-05,
+      "loss": 0.0116,
+      "step": 340
+    },
+    {
+      "epoch": 8.14,
+      "grad_norm": 0.026530884206295013,
+      "learning_rate": 9.103531961664118e-06,
+      "loss": 0.0053,
+      "step": 350
+    },
+    {
+      "epoch": 8.14,
+      "eval_loss": 0.11503978818655014,
+      "eval_runtime": 2.2223,
+      "eval_samples_per_second": 80.998,
+      "eval_steps_per_second": 3.6,
+      "step": 350
+    },
+    {
+      "epoch": 8.37,
+      "grad_norm": 0.2660503685474396,
+      "learning_rate": 7.0215196506399515e-06,
+      "loss": 0.0037,
+      "step": 360
+    },
+    {
+      "epoch": 8.6,
+      "grad_norm": 0.5767481327056885,
+      "learning_rate": 5.191722151947226e-06,
+      "loss": 0.0034,
+      "step": 370
+    },
+    {
+      "epoch": 8.84,
+      "grad_norm": 0.9280475974082947,
+      "learning_rate": 3.6248774448952695e-06,
+      "loss": 0.004,
+      "step": 380
+    },
+    {
+      "epoch": 9.07,
+      "grad_norm": 0.0765455961227417,
+      "learning_rate": 2.330180397253473e-06,
+      "loss": 0.0032,
+      "step": 390
+    },
+    {
+      "epoch": 9.3,
+      "grad_norm": 0.05842304602265358,
+      "learning_rate": 1.3152288061110518e-06,
+      "loss": 0.0021,
+      "step": 400
+    },
+    {
+      "epoch": 9.3,
+      "eval_loss": 0.13117723166942596,
+      "eval_runtime": 2.2239,
+      "eval_samples_per_second": 80.939,
+      "eval_steps_per_second": 3.597,
+      "step": 400
+    },
+    {
+      "epoch": 9.53,
+      "grad_norm": 0.02898208238184452,
+      "learning_rate": 5.859788109825793e-07,
+      "loss": 0.0026,
+      "step": 410
+    },
+    {
+      "epoch": 9.77,
+      "grad_norm": 0.41006436944007874,
+      "learning_rate": 1.4670994081297795e-07,
+      "loss": 0.0027,
+      "step": 420
+    },
+    {
+      "epoch": 10.0,
+      "grad_norm": 0.015619861893355846,
+      "learning_rate": 0.0,
+      "loss": 0.0016,
+      "step": 430
+    },
+    {
+      "epoch": 10.0,
+      "step": 430,
+      "total_flos": 2.0536791678110925e+17,
+      "train_loss": 0.0926927653899373,
+      "train_runtime": 420.9269,
+      "train_samples_per_second": 24.232,
+      "train_steps_per_second": 1.022
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 430,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 500,
+  "total_flos": 2.0536791678110925e+17,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

llama2_13b_peft/logical_deduction/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b9126db8bbe3c6a12001b7b0a43c2c93caf2551adcba85b55ad121e49d0420ac
+size 5112

llama2_13b_peft/news_commentary_es/README.md

@@ -0,0 +1,84 @@
+---
+license: other
+library_name: peft
+tags:
+- llama-factory
+- lora
+- generated_from_trainer
+base_model: /data1/model/llama2/meta-llama/Llama2-13b
+model-index:
+- name: news_commentary_es
+  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. -->
+
+# news_commentary_es
+
+This model is a fine-tuned version of [/data1/model/llama2/meta-llama/Llama2-13b](https://huggingface.co//data1/model/llama2/meta-llama/Llama2-13b) on the news_commentary_es dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.5704
+
+## 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: 1e-05
+- train_batch_size: 16
+- eval_batch_size: 16
+- seed: 42
+- distributed_type: multi-GPU
+- num_devices: 2
+- total_train_batch_size: 32
+- total_eval_batch_size: 32
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: cosine
+- lr_scheduler_warmup_steps: 20
+- num_epochs: 5.0
+
+### Training results
+
+| Training Loss | Epoch | Step | Validation Loss |
+|:-------------:|:-----:|:----:|:---------------:|
+| 0.6296        | 0.25  | 200  | 0.6344          |
+| 0.6272        | 0.5   | 400  | 0.6124          |
+| 0.6167        | 0.75  | 600  | 0.6020          |
+| 0.5882        | 1.0   | 800  | 0.5940          |
+| 0.5522        | 1.25  | 1000 | 0.5889          |
+| 0.5919        | 1.51  | 1200 | 0.5836          |
+| 0.6038        | 1.76  | 1400 | 0.5801          |
+| 0.5882        | 2.01  | 1600 | 0.5773          |
+| 0.5302        | 2.26  | 1800 | 0.5754          |
+| 0.5331        | 2.51  | 2000 | 0.5732          |
+| 0.5669        | 2.76  | 2200 | 0.5714          |
+| 0.5551        | 3.01  | 2400 | 0.5706          |
+| 0.5499        | 3.26  | 2600 | 0.5716          |
+| 0.545         | 3.51  | 2800 | 0.5704          |
+| 0.5253        | 3.76  | 3000 | 0.5704          |
+| 0.5418        | 4.02  | 3200 | 0.5702          |
+| 0.5289        | 4.27  | 3400 | 0.5704          |
+| 0.4983        | 4.52  | 3600 | 0.5706          |
+| 0.5371        | 4.77  | 3800 | 0.5705          |
+
+
+### Framework versions
+
+- PEFT 0.9.0
+- Transformers 4.38.2
+- Pytorch 2.2.1
+- Datasets 2.18.0
+- Tokenizers 0.15.2

llama2_13b_peft/news_commentary_es/adapter_config.json

@@ -0,0 +1,33 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data1/model/llama2/meta-llama/Llama2-13b",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_dropout": 0.0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "q_proj",
+    "k_proj",
+    "gate_proj",
+    "down_proj",
+    "o_proj",
+    "up_proj",
+    "v_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

llama2_13b_peft/news_commentary_es/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e8a2b2fd9e200ef442df6c06579336a490267572c4c62ea475e8fd466214a8bf
+size 125248064

llama2_13b_peft/news_commentary_es/all_results.json

@@ -0,0 +1,11 @@
+{
+    "epoch": 5.0,
+    "eval_loss": 0.5704381465911865,
+    "eval_runtime": 68.1349,
+    "eval_samples_per_second": 66.045,
+    "eval_steps_per_second": 2.069,
+    "train_loss": 0.569244503227052,
+    "train_runtime": 8741.9073,
+    "train_samples_per_second": 14.585,
+    "train_steps_per_second": 0.456
+}

llama2_13b_peft/news_commentary_es/eval_results.json

@@ -0,0 +1,7 @@
+{
+    "epoch": 5.0,
+    "eval_loss": 0.5704381465911865,
+    "eval_runtime": 68.1349,
+    "eval_samples_per_second": 66.045,
+    "eval_steps_per_second": 2.069
+}

llama2_13b_peft/news_commentary_es/special_tokens_map.json

@@ -0,0 +1,24 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "</s>",
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

llama2_13b_peft/news_commentary_es/tokenizer.model

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

llama2_13b_peft/news_commentary_es/tokenizer_config.json

@@ -0,0 +1,45 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "add_prefix_space": true,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "chat_template": "{% if messages[0]['role'] == 'system' %}{% set system_message = messages[0]['content'] %}{% endif %}{% if system_message is defined %}{{ system_message }}{% endif %}{% for message in messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ content }}{% elif message['role'] == 'assistant' %}{{ content + '\\n' }}{% endif %}{% endfor %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "</s>",
+  "legacy": true,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "</s>",
+  "padding_side": "right",
+  "sp_model_kwargs": {},
+  "spaces_between_special_tokens": false,
+  "split_special_tokens": false,
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": false
+}

llama2_13b_peft/news_commentary_es/train_results.json

@@ -0,0 +1,7 @@
+{
+    "epoch": 5.0,
+    "train_loss": 0.569244503227052,
+    "train_runtime": 8741.9073,
+    "train_samples_per_second": 14.585,
+    "train_steps_per_second": 0.456
+}

llama2_13b_peft/news_commentary_es/trainer_state.json

@@ -0,0 +1,2968 @@
+{
+  "best_metric": 0.5704381465911865,
+  "best_model_checkpoint": "ckpt/llama2_13b_fuze30_no_sys/news_commentary_es/checkpoint-3000",
+  "epoch": 5.0,
+  "eval_steps": 200,
+  "global_step": 3985,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01,
+      "grad_norm": 0.5029562711715698,
+      "learning_rate": 5e-06,
+      "loss": 1.2458,
+      "step": 10
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 0.6982043981552124,
+      "learning_rate": 1e-05,
+      "loss": 1.266,
+      "step": 20
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.7867264151573181,
+      "learning_rate": 9.99984305369385e-06,
+      "loss": 1.2072,
+      "step": 30
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.7258906364440918,
+      "learning_rate": 9.999372224628255e-06,
+      "loss": 1.0755,
+      "step": 40
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.6007847189903259,
+      "learning_rate": 9.998587542361168e-06,
+      "loss": 0.9002,
+      "step": 50
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.3854537606239319,
+      "learning_rate": 9.997489056153784e-06,
+      "loss": 0.7568,
+      "step": 60
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.3243318796157837,
+      "learning_rate": 9.996076834967442e-06,
+      "loss": 0.7157,
+      "step": 70
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.24546250700950623,
+      "learning_rate": 9.994350967459303e-06,
+      "loss": 0.6984,
+      "step": 80
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.34287428855895996,
+      "learning_rate": 9.99231156197678e-06,
+      "loss": 0.6996,
+      "step": 90
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.3073829412460327,
+      "learning_rate": 9.989958746550732e-06,
+      "loss": 0.7129,
+      "step": 100
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.2948177456855774,
+      "learning_rate": 9.987292668887439e-06,
+      "loss": 0.6977,
+      "step": 110
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 0.3140040934085846,
+      "learning_rate": 9.984313496359317e-06,
+      "loss": 0.6272,
+      "step": 120
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.2903355658054352,
+      "learning_rate": 9.981021415994413e-06,
+      "loss": 0.6285,
+      "step": 130
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 0.3696918189525604,
+      "learning_rate": 9.97741663446467e-06,
+      "loss": 0.651,
+      "step": 140
+    },
+    {
+      "epoch": 0.19,
+      "grad_norm": 0.24330583214759827,
+      "learning_rate": 9.973499378072947e-06,
+      "loss": 0.6376,
+      "step": 150
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.26600131392478943,
+      "learning_rate": 9.96926989273881e-06,
+      "loss": 0.6402,
+      "step": 160
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 0.34049201011657715,
+      "learning_rate": 9.9647284439831e-06,
+      "loss": 0.6662,
+      "step": 170
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 0.3909376859664917,
+      "learning_rate": 9.959875316911261e-06,
+      "loss": 0.6134,
+      "step": 180
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.3953985869884491,
+      "learning_rate": 9.95471081619544e-06,
+      "loss": 0.6054,
+      "step": 190
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.42383915185928345,
+      "learning_rate": 9.949235266055356e-06,
+      "loss": 0.6296,
+      "step": 200
+    },
+    {
+      "epoch": 0.25,
+      "eval_loss": 0.6343660950660706,
+      "eval_runtime": 71.7569,
+      "eval_samples_per_second": 62.712,
+      "eval_steps_per_second": 1.965,
+      "step": 200
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 0.37584221363067627,
+      "learning_rate": 9.943449010237965e-06,
+      "loss": 0.6724,
+      "step": 210
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 0.368352472782135,
+      "learning_rate": 9.937352411995851e-06,
+      "loss": 0.6065,
+      "step": 220
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 0.30686938762664795,
+      "learning_rate": 9.930945854064447e-06,
+      "loss": 0.5998,
+      "step": 230
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 0.2913750410079956,
+      "learning_rate": 9.924229738637992e-06,
+      "loss": 0.6274,
+      "step": 240
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 0.30744606256484985,
+      "learning_rate": 9.91720448734429e-06,
+      "loss": 0.6442,
+      "step": 250
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 0.315236359834671,
+      "learning_rate": 9.909870541218238e-06,
+      "loss": 0.6026,
+      "step": 260
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 0.28183722496032715,
+      "learning_rate": 9.902228360674137e-06,
+      "loss": 0.6511,
+      "step": 270
+    },
+    {
+      "epoch": 0.35,
+      "grad_norm": 0.33093082904815674,
+      "learning_rate": 9.89427842547679e-06,
+      "loss": 0.6575,
+      "step": 280
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.3711361587047577,
+      "learning_rate": 9.886021234711382e-06,
+      "loss": 0.6209,
+      "step": 290
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 0.35551759600639343,
+      "learning_rate": 9.877457306752149e-06,
+      "loss": 0.6265,
+      "step": 300
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 0.3064817786216736,
+      "learning_rate": 9.868587179229834e-06,
+      "loss": 0.6214,
+      "step": 310
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.4653981924057007,
+      "learning_rate": 9.859411408997939e-06,
+      "loss": 0.6234,
+      "step": 320
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 0.3303908407688141,
+      "learning_rate": 9.84993057209776e-06,
+      "loss": 0.6502,
+      "step": 330
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 0.3864089548587799,
+      "learning_rate": 9.84014526372223e-06,
+      "loss": 0.6275,
+      "step": 340
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 0.3393557071685791,
+      "learning_rate": 9.83005609817855e-06,
+      "loss": 0.5913,
+      "step": 350
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 0.3689006567001343,
+      "learning_rate": 9.819663708849624e-06,
+      "loss": 0.5962,
+      "step": 360
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 0.32422763109207153,
+      "learning_rate": 9.808968748154302e-06,
+      "loss": 0.5896,
+      "step": 370
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.3583005964756012,
+      "learning_rate": 9.797971887506413e-06,
+      "loss": 0.617,
+      "step": 380
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 0.331403911113739,
+      "learning_rate": 9.786673817272621e-06,
+      "loss": 0.5994,
+      "step": 390
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 0.3698234558105469,
+      "learning_rate": 9.77507524672908e-06,
+      "loss": 0.6272,
+      "step": 400
+    },
+    {
+      "epoch": 0.5,
+      "eval_loss": 0.612372875213623,
+      "eval_runtime": 68.9986,
+      "eval_samples_per_second": 65.219,
+      "eval_steps_per_second": 2.044,
+      "step": 400
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 0.3716101348400116,
+      "learning_rate": 9.763176904016914e-06,
+      "loss": 0.6246,
+      "step": 410
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 0.4471350610256195,
+      "learning_rate": 9.750979536096495e-06,
+      "loss": 0.645,
+      "step": 420
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 0.31366682052612305,
+      "learning_rate": 9.738483908700563e-06,
+      "loss": 0.6397,
+      "step": 430
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 0.3895745575428009,
+      "learning_rate": 9.72569080628614e-06,
+      "loss": 0.6118,
+      "step": 440
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.42779144644737244,
+      "learning_rate": 9.71260103198529e-06,
+      "loss": 0.6359,
+      "step": 450
+    },
+    {
+      "epoch": 0.58,
+      "grad_norm": 0.37752124667167664,
+      "learning_rate": 9.699215407554712e-06,
+      "loss": 0.5863,
+      "step": 460
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 0.43578141927719116,
+      "learning_rate": 9.685534773324124e-06,
+      "loss": 0.6244,
+      "step": 470
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 0.5181850790977478,
+      "learning_rate": 9.671559988143529e-06,
+      "loss": 0.5962,
+      "step": 480
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 0.43225836753845215,
+      "learning_rate": 9.657291929329293e-06,
+      "loss": 0.631,
+      "step": 490
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 0.36941251158714294,
+      "learning_rate": 9.64273149260907e-06,
+      "loss": 0.6398,
+      "step": 500
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.3803509473800659,
+      "learning_rate": 9.627879592065558e-06,
+      "loss": 0.6393,
+      "step": 510
+    },
+    {
+      "epoch": 0.65,
+      "grad_norm": 0.3586438298225403,
+      "learning_rate": 9.612737160079134e-06,
+      "loss": 0.5944,
+      "step": 520
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 0.41518551111221313,
+      "learning_rate": 9.597305147269303e-06,
+      "loss": 0.6224,
+      "step": 530
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.3281481862068176,
+      "learning_rate": 9.581584522435025e-06,
+      "loss": 0.5853,
+      "step": 540
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 0.3723869323730469,
+      "learning_rate": 9.565576272493902e-06,
+      "loss": 0.591,
+      "step": 550
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 0.37543514370918274,
+      "learning_rate": 9.549281402420212e-06,
+      "loss": 0.5883,
+      "step": 560
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.39954090118408203,
+      "learning_rate": 9.532700935181822e-06,
+      "loss": 0.5942,
+      "step": 570
+    },
+    {
+      "epoch": 0.73,
+      "grad_norm": 0.4060945212841034,
+      "learning_rate": 9.515835911675968e-06,
+      "loss": 0.6082,
+      "step": 580
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 0.32310950756073,
+      "learning_rate": 9.498687390663905e-06,
+      "loss": 0.5836,
+      "step": 590
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 0.37066149711608887,
+      "learning_rate": 9.481256448704445e-06,
+      "loss": 0.6167,
+      "step": 600
+    },
+    {
+      "epoch": 0.75,
+      "eval_loss": 0.6020291447639465,
+      "eval_runtime": 71.3713,
+      "eval_samples_per_second": 63.051,
+      "eval_steps_per_second": 1.976,
+      "step": 600
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 0.38338422775268555,
+      "learning_rate": 9.463544180086368e-06,
+      "loss": 0.6211,
+      "step": 610
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 0.3963073790073395,
+      "learning_rate": 9.445551696759732e-06,
+      "loss": 0.5827,
+      "step": 620
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 0.4199773967266083,
+      "learning_rate": 9.427280128266049e-06,
+      "loss": 0.5782,
+      "step": 630
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.39564579725265503,
+      "learning_rate": 9.4087306216674e-06,
+      "loss": 0.6028,
+      "step": 640
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 0.43293875455856323,
+      "learning_rate": 9.389904341474394e-06,
+      "loss": 0.5791,
+      "step": 650
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 0.3330543041229248,
+      "learning_rate": 9.370802469573088e-06,
+      "loss": 0.5875,
+      "step": 660
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.4487074613571167,
+      "learning_rate": 9.351426205150778e-06,
+      "loss": 0.6148,
+      "step": 670
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 0.4278540015220642,
+      "learning_rate": 9.331776764620711e-06,
+      "loss": 0.6271,
+      "step": 680
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 0.4620065987110138,
+      "learning_rate": 9.311855381545735e-06,
+      "loss": 0.6439,
+      "step": 690
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.5026012063026428,
+      "learning_rate": 9.291663306560843e-06,
+      "loss": 0.614,
+      "step": 700
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 0.3847663700580597,
+      "learning_rate": 9.271201807294667e-06,
+      "loss": 0.6058,
+      "step": 710
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 0.38854289054870605,
+      "learning_rate": 9.250472168289901e-06,
+      "loss": 0.5859,
+      "step": 720
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 0.4080798625946045,
+      "learning_rate": 9.22947569092265e-06,
+      "loss": 0.5874,
+      "step": 730
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 0.587581992149353,
+      "learning_rate": 9.20821369332074e-06,
+      "loss": 0.6077,
+      "step": 740
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 0.5180934071540833,
+      "learning_rate": 9.186687510280968e-06,
+      "loss": 0.6005,
+      "step": 750
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 0.4063900411128998,
+      "learning_rate": 9.164898493185295e-06,
+      "loss": 0.6007,
+      "step": 760
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 0.3806501626968384,
+      "learning_rate": 9.142848009916025e-06,
+      "loss": 0.6,
+      "step": 770
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 0.4454844892024994,
+      "learning_rate": 9.120537444769912e-06,
+      "loss": 0.5704,
+      "step": 780
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 0.4750034511089325,
+      "learning_rate": 9.097968198371277e-06,
+      "loss": 0.626,
+      "step": 790
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.44586676359176636,
+      "learning_rate": 9.075141687584056e-06,
+      "loss": 0.5882,
+      "step": 800
+    },
+    {
+      "epoch": 1.0,
+      "eval_loss": 0.5939783453941345,
+      "eval_runtime": 72.333,
+      "eval_samples_per_second": 62.212,
+      "eval_steps_per_second": 1.949,
+      "step": 800
+    },
+    {
+      "epoch": 1.02,
+      "grad_norm": 0.4351860284805298,
+      "learning_rate": 9.052059345422874e-06,
+      "loss": 0.5908,
+      "step": 810
+    },
+    {
+      "epoch": 1.03,
+      "grad_norm": 0.5268151760101318,
+      "learning_rate": 9.028722620963066e-06,
+      "loss": 0.6155,
+      "step": 820
+    },
+    {
+      "epoch": 1.04,
+      "grad_norm": 0.5248316526412964,
+      "learning_rate": 9.00513297924971e-06,
+      "loss": 0.5736,
+      "step": 830
+    },
+    {
+      "epoch": 1.05,
+      "grad_norm": 0.43092867732048035,
+      "learning_rate": 8.98129190120566e-06,
+      "loss": 0.5602,
+      "step": 840
+    },
+    {
+      "epoch": 1.07,
+      "grad_norm": 0.4282706379890442,
+      "learning_rate": 8.957200883538568e-06,
+      "loss": 0.5737,
+      "step": 850
+    },
+    {
+      "epoch": 1.08,
+      "grad_norm": 0.6257330775260925,
+      "learning_rate": 8.932861438646929e-06,
+      "loss": 0.5982,
+      "step": 860
+    },
+    {
+      "epoch": 1.09,
+      "grad_norm": 0.44131189584732056,
+      "learning_rate": 8.908275094525132e-06,
+      "loss": 0.6144,
+      "step": 870
+    },
+    {
+      "epoch": 1.1,
+      "grad_norm": 0.49996450543403625,
+      "learning_rate": 8.88344339466753e-06,
+      "loss": 0.6138,
+      "step": 880
+    },
+    {
+      "epoch": 1.12,
+      "grad_norm": 0.41754284501075745,
+      "learning_rate": 8.858367897971555e-06,
+      "loss": 0.5641,
+      "step": 890
+    },
+    {
+      "epoch": 1.13,
+      "grad_norm": 0.5091477036476135,
+      "learning_rate": 8.833050178639835e-06,
+      "loss": 0.5455,
+      "step": 900
+    },
+    {
+      "epoch": 1.14,
+      "grad_norm": 0.5106736421585083,
+      "learning_rate": 8.807491826081384e-06,
+      "loss": 0.5913,
+      "step": 910
+    },
+    {
+      "epoch": 1.15,
+      "grad_norm": 0.48033618927001953,
+      "learning_rate": 8.781694444811813e-06,
+      "loss": 0.5773,
+      "step": 920
+    },
+    {
+      "epoch": 1.17,
+      "grad_norm": 0.5638559460639954,
+      "learning_rate": 8.755659654352599e-06,
+      "loss": 0.5732,
+      "step": 930
+    },
+    {
+      "epoch": 1.18,
+      "grad_norm": 0.45605936646461487,
+      "learning_rate": 8.729389089129423e-06,
+      "loss": 0.5911,
+      "step": 940
+    },
+    {
+      "epoch": 1.19,
+      "grad_norm": 0.4803718328475952,
+      "learning_rate": 8.702884398369552e-06,
+      "loss": 0.5909,
+      "step": 950
+    },
+    {
+      "epoch": 1.2,
+      "grad_norm": 0.5215190649032593,
+      "learning_rate": 8.676147245998307e-06,
+      "loss": 0.614,
+      "step": 960
+    },
+    {
+      "epoch": 1.22,
+      "grad_norm": 0.490953654050827,
+      "learning_rate": 8.649179310534612e-06,
+      "loss": 0.5687,
+      "step": 970
+    },
+    {
+      "epoch": 1.23,
+      "grad_norm": 0.5034527778625488,
+      "learning_rate": 8.621982284985611e-06,
+      "loss": 0.5948,
+      "step": 980
+    },
+    {
+      "epoch": 1.24,
+      "grad_norm": 0.5091165900230408,
+      "learning_rate": 8.59455787674038e-06,
+      "loss": 0.5695,
+      "step": 990
+    },
+    {
+      "epoch": 1.25,
+      "grad_norm": 0.49061551690101624,
+      "learning_rate": 8.56690780746275e-06,
+      "loss": 0.5522,
+      "step": 1000
+    },
+    {
+      "epoch": 1.25,
+      "eval_loss": 0.5888673067092896,
+      "eval_runtime": 73.1303,
+      "eval_samples_per_second": 61.534,
+      "eval_steps_per_second": 1.928,
+      "step": 1000
+    },
+    {
+      "epoch": 1.27,
+      "grad_norm": 0.4834423065185547,
+      "learning_rate": 8.539033812983215e-06,
+      "loss": 0.5413,
+      "step": 1010
+    },
+    {
+      "epoch": 1.28,
+      "grad_norm": 0.6021168828010559,
+      "learning_rate": 8.510937643189962e-06,
+      "loss": 0.5882,
+      "step": 1020
+    },
+    {
+      "epoch": 1.29,
+      "grad_norm": 0.5302062630653381,
+      "learning_rate": 8.482621061919021e-06,
+      "loss": 0.6059,
+      "step": 1030
+    },
+    {
+      "epoch": 1.3,
+      "grad_norm": 0.4995095729827881,
+      "learning_rate": 8.454085846843523e-06,
+      "loss": 0.5419,
+      "step": 1040
+    },
+    {
+      "epoch": 1.32,
+      "grad_norm": 0.5463216304779053,
+      "learning_rate": 8.425333789362108e-06,
+      "loss": 0.5941,
+      "step": 1050
+    },
+    {
+      "epoch": 1.33,
+      "grad_norm": 0.536431610584259,
+      "learning_rate": 8.396366694486466e-06,
+      "loss": 0.6284,
+      "step": 1060
+    },
+    {
+      "epoch": 1.34,
+      "grad_norm": 0.5217477083206177,
+      "learning_rate": 8.36718638072801e-06,
+      "loss": 0.6178,
+      "step": 1070
+    },
+    {
+      "epoch": 1.36,
+      "grad_norm": 0.4792711138725281,
+      "learning_rate": 8.337794679983724e-06,
+      "loss": 0.5858,
+      "step": 1080
+    },
+    {
+      "epoch": 1.37,
+      "grad_norm": 0.5548479557037354,
+      "learning_rate": 8.30819343742115e-06,
+      "loss": 0.5861,
+      "step": 1090
+    },
+    {
+      "epoch": 1.38,
+      "grad_norm": 0.4717773199081421,
+      "learning_rate": 8.278384511362562e-06,
+      "loss": 0.6047,
+      "step": 1100
+    },
+    {
+      "epoch": 1.39,
+      "grad_norm": 0.5312036275863647,
+      "learning_rate": 8.248369773168297e-06,
+      "loss": 0.5994,
+      "step": 1110
+    },
+    {
+      "epoch": 1.41,
+      "grad_norm": 0.4545268714427948,
+      "learning_rate": 8.218151107119266e-06,
+      "loss": 0.5618,
+      "step": 1120
+    },
+    {
+      "epoch": 1.42,
+      "grad_norm": 0.6215177774429321,
+      "learning_rate": 8.187730410298675e-06,
+      "loss": 0.6228,
+      "step": 1130
+    },
+    {
+      "epoch": 1.43,
+      "grad_norm": 0.5464479923248291,
+      "learning_rate": 8.157109592472924e-06,
+      "loss": 0.5774,
+      "step": 1140
+    },
+    {
+      "epoch": 1.44,
+      "grad_norm": 0.5225650668144226,
+      "learning_rate": 8.126290575971714e-06,
+      "loss": 0.5804,
+      "step": 1150
+    },
+    {
+      "epoch": 1.46,
+      "grad_norm": 0.6188873648643494,
+      "learning_rate": 8.095275295567361e-06,
+      "loss": 0.5498,
+      "step": 1160
+    },
+    {
+      "epoch": 1.47,
+      "grad_norm": 0.5218258500099182,
+      "learning_rate": 8.064065698353344e-06,
+      "loss": 0.5829,
+      "step": 1170
+    },
+    {
+      "epoch": 1.48,
+      "grad_norm": 0.5987175703048706,
+      "learning_rate": 8.032663743622065e-06,
+      "loss": 0.5914,
+      "step": 1180
+    },
+    {
+      "epoch": 1.49,
+      "grad_norm": 0.527808666229248,
+      "learning_rate": 8.001071402741843e-06,
+      "loss": 0.5682,
+      "step": 1190
+    },
+    {
+      "epoch": 1.51,
+      "grad_norm": 0.5273358821868896,
+      "learning_rate": 7.969290659033156e-06,
+      "loss": 0.5919,
+      "step": 1200
+    },
+    {
+      "epoch": 1.51,
+      "eval_loss": 0.5836370587348938,
+      "eval_runtime": 71.4972,
+      "eval_samples_per_second": 62.939,
+      "eval_steps_per_second": 1.972,
+      "step": 1200
+    },
+    {
+      "epoch": 1.52,
+      "grad_norm": 0.6059889197349548,
+      "learning_rate": 7.93732350764414e-06,
+      "loss": 0.5763,
+      "step": 1210
+    },
+    {
+      "epoch": 1.53,
+      "grad_norm": 0.5306861996650696,
+      "learning_rate": 7.905171955425328e-06,
+      "loss": 0.5567,
+      "step": 1220
+    },
+    {
+      "epoch": 1.54,
+      "grad_norm": 0.5967408418655396,
+      "learning_rate": 7.872838020803662e-06,
+      "loss": 0.5609,
+      "step": 1230
+    },
+    {
+      "epoch": 1.56,
+      "grad_norm": 0.6726911067962646,
+      "learning_rate": 7.84032373365578e-06,
+      "loss": 0.6471,
+      "step": 1240
+    },
+    {
+      "epoch": 1.57,
+      "grad_norm": 0.5502099990844727,
+      "learning_rate": 7.807631135180591e-06,
+      "loss": 0.5886,
+      "step": 1250
+    },
+    {
+      "epoch": 1.58,
+      "grad_norm": 0.615433394908905,
+      "learning_rate": 7.774762277771123e-06,
+      "loss": 0.5559,
+      "step": 1260
+    },
+    {
+      "epoch": 1.59,
+      "grad_norm": 0.5481686592102051,
+      "learning_rate": 7.741719224885675e-06,
+      "loss": 0.5928,
+      "step": 1270
+    },
+    {
+      "epoch": 1.61,
+      "grad_norm": 0.620995044708252,
+      "learning_rate": 7.708504050918288e-06,
+      "loss": 0.5683,
+      "step": 1280
+    },
+    {
+      "epoch": 1.62,
+      "grad_norm": 0.5698446035385132,
+      "learning_rate": 7.675118841068507e-06,
+      "loss": 0.5909,
+      "step": 1290
+    },
+    {
+      "epoch": 1.63,
+      "grad_norm": 0.5968176126480103,
+      "learning_rate": 7.64156569121048e-06,
+      "loss": 0.573,
+      "step": 1300
+    },
+    {
+      "epoch": 1.64,
+      "grad_norm": 0.7233667373657227,
+      "learning_rate": 7.6078467077613746e-06,
+      "loss": 0.5888,
+      "step": 1310
+    },
+    {
+      "epoch": 1.66,
+      "grad_norm": 0.5293125510215759,
+      "learning_rate": 7.5739640075491546e-06,
+      "loss": 0.6316,
+      "step": 1320
+    },
+    {
+      "epoch": 1.67,
+      "grad_norm": 0.5775229334831238,
+      "learning_rate": 7.5399197176796735e-06,
+      "loss": 0.591,
+      "step": 1330
+    },
+    {
+      "epoch": 1.68,
+      "grad_norm": 0.7166782021522522,
+      "learning_rate": 7.50571597540315e-06,
+      "loss": 0.5505,
+      "step": 1340
+    },
+    {
+      "epoch": 1.69,
+      "grad_norm": 0.5287649631500244,
+      "learning_rate": 7.471354927979984e-06,
+      "loss": 0.6036,
+      "step": 1350
+    },
+    {
+      "epoch": 1.71,
+      "grad_norm": 0.5856715440750122,
+      "learning_rate": 7.436838732545964e-06,
+      "loss": 0.5682,
+      "step": 1360
+    },
+    {
+      "epoch": 1.72,
+      "grad_norm": 0.5226353406906128,
+      "learning_rate": 7.402169555976841e-06,
+      "loss": 0.5593,
+      "step": 1370
+    },
+    {
+      "epoch": 1.73,
+      "grad_norm": 0.6625097393989563,
+      "learning_rate": 7.367349574752296e-06,
+      "loss": 0.5853,
+      "step": 1380
+    },
+    {
+      "epoch": 1.74,
+      "grad_norm": 0.5979090929031372,
+      "learning_rate": 7.332380974819301e-06,
+      "loss": 0.5493,
+      "step": 1390
+    },
+    {
+      "epoch": 1.76,
+      "grad_norm": 0.6176837682723999,
+      "learning_rate": 7.297265951454891e-06,
+      "loss": 0.6038,
+      "step": 1400
+    },
+    {
+      "epoch": 1.76,
+      "eval_loss": 0.5800889730453491,
+      "eval_runtime": 69.6671,
+      "eval_samples_per_second": 64.593,
+      "eval_steps_per_second": 2.024,
+      "step": 1400
+    },
+    {
+      "epoch": 1.77,
+      "grad_norm": 0.6878402233123779,
+      "learning_rate": 7.262006709128351e-06,
+      "loss": 0.5453,
+      "step": 1410
+    },
+    {
+      "epoch": 1.78,
+      "grad_norm": 0.6541485786437988,
+      "learning_rate": 7.226605461362817e-06,
+      "loss": 0.586,
+      "step": 1420
+    },
+    {
+      "epoch": 1.79,
+      "grad_norm": 0.6143850088119507,
+      "learning_rate": 7.191064430596315e-06,
+      "loss": 0.5422,
+      "step": 1430
+    },
+    {
+      "epoch": 1.81,
+      "grad_norm": 0.6676700115203857,
+      "learning_rate": 7.155385848042245e-06,
+      "loss": 0.5806,
+      "step": 1440
+    },
+    {
+      "epoch": 1.82,
+      "grad_norm": 0.6332250833511353,
+      "learning_rate": 7.119571953549305e-06,
+      "loss": 0.5585,
+      "step": 1450
+    },
+    {
+      "epoch": 1.83,
+      "grad_norm": 0.7023981809616089,
+      "learning_rate": 7.08362499546087e-06,
+      "loss": 0.5833,
+      "step": 1460
+    },
+    {
+      "epoch": 1.84,
+      "grad_norm": 0.665985107421875,
+      "learning_rate": 7.047547230473861e-06,
+      "loss": 0.569,
+      "step": 1470
+    },
+    {
+      "epoch": 1.86,
+      "grad_norm": 0.5501468777656555,
+      "learning_rate": 7.011340923497055e-06,
+      "loss": 0.5824,
+      "step": 1480
+    },
+    {
+      "epoch": 1.87,
+      "grad_norm": 0.5913044810295105,
+      "learning_rate": 6.975008347508906e-06,
+      "loss": 0.5581,
+      "step": 1490
+    },
+    {
+      "epoch": 1.88,
+      "grad_norm": 0.6394723653793335,
+      "learning_rate": 6.938551783414856e-06,
+      "loss": 0.5813,
+      "step": 1500
+    },
+    {
+      "epoch": 1.89,
+      "grad_norm": 0.6261583566665649,
+      "learning_rate": 6.901973519904128e-06,
+      "loss": 0.5804,
+      "step": 1510
+    },
+    {
+      "epoch": 1.91,
+      "grad_norm": 0.5661801695823669,
+      "learning_rate": 6.865275853306062e-06,
+      "loss": 0.5587,
+      "step": 1520
+    },
+    {
+      "epoch": 1.92,
+      "grad_norm": 0.5941905975341797,
+      "learning_rate": 6.828461087445946e-06,
+      "loss": 0.5735,
+      "step": 1530
+    },
+    {
+      "epoch": 1.93,
+      "grad_norm": 0.5802645087242126,
+      "learning_rate": 6.791531533500384e-06,
+      "loss": 0.5854,
+      "step": 1540
+    },
+    {
+      "epoch": 1.94,
+      "grad_norm": 0.5715972781181335,
+      "learning_rate": 6.754489509852208e-06,
+      "loss": 0.5627,
+      "step": 1550
+    },
+    {
+      "epoch": 1.96,
+      "grad_norm": 0.6669459342956543,
+      "learning_rate": 6.7173373419449295e-06,
+      "loss": 0.5647,
+      "step": 1560
+    },
+    {
+      "epoch": 1.97,
+      "grad_norm": 0.6907923221588135,
+      "learning_rate": 6.68007736213676e-06,
+      "loss": 0.5364,
+      "step": 1570
+    },
+    {
+      "epoch": 1.98,
+      "grad_norm": 0.5516065359115601,
+      "learning_rate": 6.6427119095541745e-06,
+      "loss": 0.5643,
+      "step": 1580
+    },
+    {
+      "epoch": 1.99,
+      "grad_norm": 0.6715050935745239,
+      "learning_rate": 6.60524332994508e-06,
+      "loss": 0.6026,
+      "step": 1590
+    },
+    {
+      "epoch": 2.01,
+      "grad_norm": 0.5935114026069641,
+      "learning_rate": 6.567673975531542e-06,
+      "loss": 0.5882,
+      "step": 1600
+    },
+    {
+      "epoch": 2.01,
+      "eval_loss": 0.5773123502731323,
+      "eval_runtime": 71.3304,
+      "eval_samples_per_second": 63.087,
+      "eval_steps_per_second": 1.977,
+      "step": 1600
+    },
+    {
+      "epoch": 2.02,
+      "grad_norm": 0.7686386108398438,
+      "learning_rate": 6.530006204862121e-06,
+      "loss": 0.5391,
+      "step": 1610
+    },
+    {
+      "epoch": 2.03,
+      "grad_norm": 0.6634293794631958,
+      "learning_rate": 6.492242382663804e-06,
+      "loss": 0.5495,
+      "step": 1620
+    },
+    {
+      "epoch": 2.05,
+      "grad_norm": 0.8547865748405457,
+      "learning_rate": 6.454384879693549e-06,
+      "loss": 0.5456,
+      "step": 1630
+    },
+    {
+      "epoch": 2.06,
+      "grad_norm": 0.8082049489021301,
+      "learning_rate": 6.41643607258946e-06,
+      "loss": 0.5903,
+      "step": 1640
+    },
+    {
+      "epoch": 2.07,
+      "grad_norm": 0.6522872447967529,
+      "learning_rate": 6.378398343721575e-06,
+      "loss": 0.5476,
+      "step": 1650
+    },
+    {
+      "epoch": 2.08,
+      "grad_norm": 0.663383960723877,
+      "learning_rate": 6.340274081042307e-06,
+      "loss": 0.5271,
+      "step": 1660
+    },
+    {
+      "epoch": 2.1,
+      "grad_norm": 0.8022051453590393,
+      "learning_rate": 6.3020656779365416e-06,
+      "loss": 0.5656,
+      "step": 1670
+    },
+    {
+      "epoch": 2.11,
+      "grad_norm": 0.7163932919502258,
+      "learning_rate": 6.263775533071368e-06,
+      "loss": 0.5581,
+      "step": 1680
+    },
+    {
+      "epoch": 2.12,
+      "grad_norm": 0.6052596569061279,
+      "learning_rate": 6.225406050245507e-06,
+      "loss": 0.554,
+      "step": 1690
+    },
+    {
+      "epoch": 2.13,
+      "grad_norm": 0.6838263273239136,
+      "learning_rate": 6.186959638238397e-06,
+      "loss": 0.5812,
+      "step": 1700
+    },
+    {
+      "epoch": 2.15,
+      "grad_norm": 0.6921390891075134,
+      "learning_rate": 6.148438710658979e-06,
+      "loss": 0.5237,
+      "step": 1710
+    },
+    {
+      "epoch": 2.16,
+      "grad_norm": 0.883646547794342,
+      "learning_rate": 6.109845685794169e-06,
+      "loss": 0.5267,
+      "step": 1720
+    },
+    {
+      "epoch": 2.17,
+      "grad_norm": 0.6446242928504944,
+      "learning_rate": 6.071182986457046e-06,
+      "loss": 0.5466,
+      "step": 1730
+    },
+    {
+      "epoch": 2.18,
+      "grad_norm": 0.7913056015968323,
+      "learning_rate": 6.032453039834748e-06,
+      "loss": 0.573,
+      "step": 1740
+    },
+    {
+      "epoch": 2.2,
+      "grad_norm": 0.7966864705085754,
+      "learning_rate": 5.993658277336101e-06,
+      "loss": 0.5498,
+      "step": 1750
+    },
+    {
+      "epoch": 2.21,
+      "grad_norm": 0.7100770473480225,
+      "learning_rate": 5.954801134438971e-06,
+      "loss": 0.5624,
+      "step": 1760
+    },
+    {
+      "epoch": 2.22,
+      "grad_norm": 0.5544886589050293,
+      "learning_rate": 5.915884050537378e-06,
+      "loss": 0.5349,
+      "step": 1770
+    },
+    {
+      "epoch": 2.23,
+      "grad_norm": 0.7371806502342224,
+      "learning_rate": 5.8769094687883484e-06,
+      "loss": 0.5789,
+      "step": 1780
+    },
+    {
+      "epoch": 2.25,
+      "grad_norm": 0.7266082763671875,
+      "learning_rate": 5.837879835958535e-06,
+      "loss": 0.5573,
+      "step": 1790
+    },
+    {
+      "epoch": 2.26,
+      "grad_norm": 0.6665763854980469,
+      "learning_rate": 5.798797602270623e-06,
+      "loss": 0.5302,
+      "step": 1800
+    },
+    {
+      "epoch": 2.26,
+      "eval_loss": 0.5754047632217407,
+      "eval_runtime": 72.9642,
+      "eval_samples_per_second": 61.674,
+      "eval_steps_per_second": 1.932,
+      "step": 1800
+    },
+    {
+      "epoch": 2.27,
+      "grad_norm": 0.7089454531669617,
+      "learning_rate": 5.7596652212494944e-06,
+      "loss": 0.5652,
+      "step": 1810
+    },
+    {
+      "epoch": 2.28,
+      "grad_norm": 0.790558397769928,
+      "learning_rate": 5.720485149568211e-06,
+      "loss": 0.5547,
+      "step": 1820
+    },
+    {
+      "epoch": 2.3,
+      "grad_norm": 0.7344884872436523,
+      "learning_rate": 5.681259846893783e-06,
+      "loss": 0.5703,
+      "step": 1830
+    },
+    {
+      "epoch": 2.31,
+      "grad_norm": 0.7799204587936401,
+      "learning_rate": 5.641991775732756e-06,
+      "loss": 0.5369,
+      "step": 1840
+    },
+    {
+      "epoch": 2.32,
+      "grad_norm": 0.7437064051628113,
+      "learning_rate": 5.6026834012766155e-06,
+      "loss": 0.5814,
+      "step": 1850
+    },
+    {
+      "epoch": 2.33,
+      "grad_norm": 0.7207692861557007,
+      "learning_rate": 5.5633371912470334e-06,
+      "loss": 0.5535,
+      "step": 1860
+    },
+    {
+      "epoch": 2.35,
+      "grad_norm": 0.7650673985481262,
+      "learning_rate": 5.5239556157409366e-06,
+      "loss": 0.5507,
+      "step": 1870
+    },
+    {
+      "epoch": 2.36,
+      "grad_norm": 0.8299598097801208,
+      "learning_rate": 5.484541147075449e-06,
+      "loss": 0.5678,
+      "step": 1880
+    },
+    {
+      "epoch": 2.37,
+      "grad_norm": 0.6018086671829224,
+      "learning_rate": 5.4450962596326776e-06,
+      "loss": 0.5749,
+      "step": 1890
+    },
+    {
+      "epoch": 2.38,
+      "grad_norm": 0.7759122848510742,
+      "learning_rate": 5.405623429704373e-06,
+      "loss": 0.5561,
+      "step": 1900
+    },
+    {
+      "epoch": 2.4,
+      "grad_norm": 0.7874650359153748,
+      "learning_rate": 5.366125135336477e-06,
+      "loss": 0.568,
+      "step": 1910
+    },
+    {
+      "epoch": 2.41,
+      "grad_norm": 0.642218828201294,
+      "learning_rate": 5.326603856173548e-06,
+      "loss": 0.5468,
+      "step": 1920
+    },
+    {
+      "epoch": 2.42,
+      "grad_norm": 0.7558509707450867,
+      "learning_rate": 5.2870620733030994e-06,
+      "loss": 0.5315,
+      "step": 1930
+    },
+    {
+      "epoch": 2.43,
+      "grad_norm": 0.846817672252655,
+      "learning_rate": 5.247502269099834e-06,
+      "loss": 0.5511,
+      "step": 1940
+    },
+    {
+      "epoch": 2.45,
+      "grad_norm": 0.725319504737854,
+      "learning_rate": 5.207926927069809e-06,
+      "loss": 0.5098,
+      "step": 1950
+    },
+    {
+      "epoch": 2.46,
+      "grad_norm": 0.7795179486274719,
+      "learning_rate": 5.168338531694523e-06,
+      "loss": 0.5417,
+      "step": 1960
+    },
+    {
+      "epoch": 2.47,
+      "grad_norm": 0.764240562915802,
+      "learning_rate": 5.1287395682749444e-06,
+      "loss": 0.5502,
+      "step": 1970
+    },
+    {
+      "epoch": 2.48,
+      "grad_norm": 0.7871197462081909,
+      "learning_rate": 5.089132522775486e-06,
+      "loss": 0.575,
+      "step": 1980
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 0.8167646527290344,
+      "learning_rate": 5.049519881667945e-06,
+      "loss": 0.5637,
+      "step": 1990
+    },
+    {
+      "epoch": 2.51,
+      "grad_norm": 0.7696388959884644,
+      "learning_rate": 5.009904131775402e-06,
+      "loss": 0.5331,
+      "step": 2000
+    },
+    {
+      "epoch": 2.51,
+      "eval_loss": 0.5732148289680481,
+      "eval_runtime": 73.9926,
+      "eval_samples_per_second": 60.817,
+      "eval_steps_per_second": 1.906,
+      "step": 2000
+    },
+    {
+      "epoch": 2.52,
+      "grad_norm": 0.8611487746238708,
+      "learning_rate": 4.970287760116098e-06,
+      "loss": 0.544,
+      "step": 2010
+    },
+    {
+      "epoch": 2.53,
+      "grad_norm": 0.7960777878761292,
+      "learning_rate": 4.930673253747314e-06,
+      "loss": 0.5702,
+      "step": 2020
+    },
+    {
+      "epoch": 2.55,
+      "grad_norm": 0.8783027529716492,
+      "learning_rate": 4.8910630996092265e-06,
+      "loss": 0.5309,
+      "step": 2030
+    },
+    {
+      "epoch": 2.56,
+      "grad_norm": 0.7091485261917114,
+      "learning_rate": 4.851459784368787e-06,
+      "loss": 0.5565,
+      "step": 2040
+    },
+    {
+      "epoch": 2.57,
+      "grad_norm": 0.7412302494049072,
+      "learning_rate": 4.811865794263611e-06,
+      "loss": 0.5546,
+      "step": 2050
+    },
+    {
+      "epoch": 2.58,
+      "grad_norm": 0.7188848257064819,
+      "learning_rate": 4.772283614945895e-06,
+      "loss": 0.5927,
+      "step": 2060
+    },
+    {
+      "epoch": 2.6,
+      "grad_norm": 0.9045892953872681,
+      "learning_rate": 4.732715731326372e-06,
+      "loss": 0.5515,
+      "step": 2070
+    },
+    {
+      "epoch": 2.61,
+      "grad_norm": 0.7781686186790466,
+      "learning_rate": 4.693164627418315e-06,
+      "loss": 0.5289,
+      "step": 2080
+    },
+    {
+      "epoch": 2.62,
+      "grad_norm": 0.8868610858917236,
+      "learning_rate": 4.653632786181587e-06,
+      "loss": 0.5934,
+      "step": 2090
+    },
+    {
+      "epoch": 2.63,
+      "grad_norm": 0.7511059641838074,
+      "learning_rate": 4.614122689366769e-06,
+      "loss": 0.5464,
+      "step": 2100
+    },
+    {
+      "epoch": 2.65,
+      "grad_norm": 0.7453913688659668,
+      "learning_rate": 4.574636817359365e-06,
+      "loss": 0.5318,
+      "step": 2110
+    },
+    {
+      "epoch": 2.66,
+      "grad_norm": 0.6865512132644653,
+      "learning_rate": 4.535177649024077e-06,
+      "loss": 0.5527,
+      "step": 2120
+    },
+    {
+      "epoch": 2.67,
+      "grad_norm": 0.8288848400115967,
+      "learning_rate": 4.495747661549187e-06,
+      "loss": 0.5368,
+      "step": 2130
+    },
+    {
+      "epoch": 2.69,
+      "grad_norm": 0.8635351657867432,
+      "learning_rate": 4.456349330291055e-06,
+      "loss": 0.5549,
+      "step": 2140
+    },
+    {
+      "epoch": 2.7,
+      "grad_norm": 0.7543601393699646,
+      "learning_rate": 4.4169851286187005e-06,
+      "loss": 0.5435,
+      "step": 2150
+    },
+    {
+      "epoch": 2.71,
+      "grad_norm": 0.8399900794029236,
+      "learning_rate": 4.3776575277585445e-06,
+      "loss": 0.5853,
+      "step": 2160
+    },
+    {
+      "epoch": 2.72,
+      "grad_norm": 0.9056334495544434,
+      "learning_rate": 4.3383689966392604e-06,
+      "loss": 0.5737,
+      "step": 2170
+    },
+    {
+      "epoch": 2.74,
+      "grad_norm": 0.7865429520606995,
+      "learning_rate": 4.2991220017367795e-06,
+      "loss": 0.5908,
+      "step": 2180
+    },
+    {
+      "epoch": 2.75,
+      "grad_norm": 0.9085346460342407,
+      "learning_rate": 4.259919006919456e-06,
+      "loss": 0.5468,
+      "step": 2190
+    },
+    {
+      "epoch": 2.76,
+      "grad_norm": 0.844716489315033,
+      "learning_rate": 4.220762473293379e-06,
+      "loss": 0.5669,
+      "step": 2200
+    },
+    {
+      "epoch": 2.76,
+      "eval_loss": 0.5714393854141235,
+      "eval_runtime": 74.9812,
+      "eval_samples_per_second": 60.015,
+      "eval_steps_per_second": 1.88,
+      "step": 2200
+    },
+    {
+      "epoch": 2.77,
+      "grad_norm": 0.8216796517372131,
+      "learning_rate": 4.181654859047873e-06,
+      "loss": 0.5516,
+      "step": 2210
+    },
+    {
+      "epoch": 2.79,
+      "grad_norm": 0.8482044339179993,
+      "learning_rate": 4.1425986193011805e-06,
+      "loss": 0.5463,
+      "step": 2220
+    },
+    {
+      "epoch": 2.8,
+      "grad_norm": 0.8037722706794739,
+      "learning_rate": 4.103596205946323e-06,
+      "loss": 0.5387,
+      "step": 2230
+    },
+    {
+      "epoch": 2.81,
+      "grad_norm": 0.8523209691047668,
+      "learning_rate": 4.0646500674971826e-06,
+      "loss": 0.5282,
+      "step": 2240
+    },
+    {
+      "epoch": 2.82,
+      "grad_norm": 0.8726175427436829,
+      "learning_rate": 4.0257626489347895e-06,
+      "loss": 0.5476,
+      "step": 2250
+    },
+    {
+      "epoch": 2.84,
+      "grad_norm": 0.9041292071342468,
+      "learning_rate": 3.986936391553822e-06,
+      "loss": 0.5677,
+      "step": 2260
+    },
+    {
+      "epoch": 2.85,
+      "grad_norm": 0.8939766883850098,
+      "learning_rate": 3.948173732809351e-06,
+      "loss": 0.5711,
+      "step": 2270
+    },
+    {
+      "epoch": 2.86,
+      "grad_norm": 0.7777726650238037,
+      "learning_rate": 3.909477106163818e-06,
+      "loss": 0.5289,
+      "step": 2280
+    },
+    {
+      "epoch": 2.87,
+      "grad_norm": 0.7699301242828369,
+      "learning_rate": 3.87084894093427e-06,
+      "loss": 0.55,
+      "step": 2290
+    },
+    {
+      "epoch": 2.89,
+      "grad_norm": 0.9905028939247131,
+      "learning_rate": 3.832291662139845e-06,
+      "loss": 0.541,
+      "step": 2300
+    },
+    {
+      "epoch": 2.9,
+      "grad_norm": 0.8651571869850159,
+      "learning_rate": 3.7938076903495363e-06,
+      "loss": 0.5533,
+      "step": 2310
+    },
+    {
+      "epoch": 2.91,
+      "grad_norm": 0.7639217376708984,
+      "learning_rate": 3.7553994415302274e-06,
+      "loss": 0.5511,
+      "step": 2320
+    },
+    {
+      "epoch": 2.92,
+      "grad_norm": 0.8610684871673584,
+      "learning_rate": 3.7170693268950354e-06,
+      "loss": 0.5119,
+      "step": 2330
+    },
+    {
+      "epoch": 2.94,
+      "grad_norm": 0.7909624576568604,
+      "learning_rate": 3.67881975275192e-06,
+      "loss": 0.5881,
+      "step": 2340
+    },
+    {
+      "epoch": 2.95,
+      "grad_norm": 0.9253331422805786,
+      "learning_rate": 3.6406531203526274e-06,
+      "loss": 0.5612,
+      "step": 2350
+    },
+    {
+      "epoch": 2.96,
+      "grad_norm": 0.8581608533859253,
+      "learning_rate": 3.6025718257419532e-06,
+      "loss": 0.5322,
+      "step": 2360
+    },
+    {
+      "epoch": 2.97,
+      "grad_norm": 0.9016821384429932,
+      "learning_rate": 3.564578259607303e-06,
+      "loss": 0.5487,
+      "step": 2370
+    },
+    {
+      "epoch": 2.99,
+      "grad_norm": 0.8990395665168762,
+      "learning_rate": 3.5266748071286205e-06,
+      "loss": 0.5471,
+      "step": 2380
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 0.9548435211181641,
+      "learning_rate": 3.488863847828652e-06,
+      "loss": 0.547,
+      "step": 2390
+    },
+    {
+      "epoch": 3.01,
+      "grad_norm": 0.771557092666626,
+      "learning_rate": 3.4511477554235516e-06,
+      "loss": 0.5551,
+      "step": 2400
+    },
+    {
+      "epoch": 3.01,
+      "eval_loss": 0.5705960392951965,
+      "eval_runtime": 76.5829,
+      "eval_samples_per_second": 58.76,
+      "eval_steps_per_second": 1.841,
+      "step": 2400
+    },
+    {
+      "epoch": 3.02,
+      "grad_norm": 0.8474746942520142,
+      "learning_rate": 3.4135288976738762e-06,
+      "loss": 0.5552,
+      "step": 2410
+    },
+    {
+      "epoch": 3.04,
+      "grad_norm": 1.0370606184005737,
+      "learning_rate": 3.3760096362359306e-06,
+      "loss": 0.5456,
+      "step": 2420
+    },
+    {
+      "epoch": 3.05,
+      "grad_norm": 0.7640136480331421,
+      "learning_rate": 3.33859232651351e-06,
+      "loss": 0.5228,
+      "step": 2430
+    },
+    {
+      "epoch": 3.06,
+      "grad_norm": 1.0315712690353394,
+      "learning_rate": 3.301279317510038e-06,
+      "loss": 0.5389,
+      "step": 2440
+    },
+    {
+      "epoch": 3.07,
+      "grad_norm": 1.0031884908676147,
+      "learning_rate": 3.2640729516810844e-06,
+      "loss": 0.5153,
+      "step": 2450
+    },
+    {
+      "epoch": 3.09,
+      "grad_norm": 0.8636874556541443,
+      "learning_rate": 3.226975564787322e-06,
+      "loss": 0.527,
+      "step": 2460
+    },
+    {
+      "epoch": 3.1,
+      "grad_norm": 0.9102957248687744,
+      "learning_rate": 3.1899894857478897e-06,
+      "loss": 0.5397,
+      "step": 2470
+    },
+    {
+      "epoch": 3.11,
+      "grad_norm": 0.8529887795448303,
+      "learning_rate": 3.15311703649418e-06,
+      "loss": 0.5184,
+      "step": 2480
+    },
+    {
+      "epoch": 3.12,
+      "grad_norm": 0.9462995529174805,
+      "learning_rate": 3.116360531824074e-06,
+      "loss": 0.5559,
+      "step": 2490
+    },
+    {
+      "epoch": 3.14,
+      "grad_norm": 0.8420120477676392,
+      "learning_rate": 3.0797222792566306e-06,
+      "loss": 0.5245,
+      "step": 2500
+    },
+    {
+      "epoch": 3.15,
+      "grad_norm": 0.7965397238731384,
+      "learning_rate": 3.043204578887209e-06,
+      "loss": 0.5585,
+      "step": 2510
+    },
+    {
+      "epoch": 3.16,
+      "grad_norm": 0.9633888602256775,
+      "learning_rate": 3.00680972324308e-06,
+      "loss": 0.4956,
+      "step": 2520
+    },
+    {
+      "epoch": 3.17,
+      "grad_norm": 1.1136802434921265,
+      "learning_rate": 2.970539997139509e-06,
+      "loss": 0.5444,
+      "step": 2530
+    },
+    {
+      "epoch": 3.19,
+      "grad_norm": 0.8797339797019958,
+      "learning_rate": 2.9343976775363097e-06,
+      "loss": 0.5253,
+      "step": 2540
+    },
+    {
+      "epoch": 3.2,
+      "grad_norm": 1.0063321590423584,
+      "learning_rate": 2.8983850333949065e-06,
+      "loss": 0.5427,
+      "step": 2550
+    },
+    {
+      "epoch": 3.21,
+      "grad_norm": 0.8839291930198669,
+      "learning_rate": 2.862504325535886e-06,
+      "loss": 0.5346,
+      "step": 2560
+    },
+    {
+      "epoch": 3.22,
+      "grad_norm": 0.9732418060302734,
+      "learning_rate": 2.8267578064970735e-06,
+      "loss": 0.5889,
+      "step": 2570
+    },
+    {
+      "epoch": 3.24,
+      "grad_norm": 1.0936757326126099,
+      "learning_rate": 2.791147720392118e-06,
+      "loss": 0.5136,
+      "step": 2580
+    },
+    {
+      "epoch": 3.25,
+      "grad_norm": 0.9694124460220337,
+      "learning_rate": 2.755676302769611e-06,
+      "loss": 0.5276,
+      "step": 2590
+    },
+    {
+      "epoch": 3.26,
+      "grad_norm": 0.9088903069496155,
+      "learning_rate": 2.7203457804727364e-06,
+      "loss": 0.5499,
+      "step": 2600
+    },
+    {
+      "epoch": 3.26,
+      "eval_loss": 0.5715650320053101,
+      "eval_runtime": 75.4356,
+      "eval_samples_per_second": 59.653,
+      "eval_steps_per_second": 1.869,
+      "step": 2600
+    },
+    {
+      "epoch": 3.27,
+      "grad_norm": 0.8832049369812012,
+      "learning_rate": 2.6851583714994843e-06,
+      "loss": 0.5392,
+      "step": 2610
+    },
+    {
+      "epoch": 3.29,
+      "grad_norm": 1.0563007593154907,
+      "learning_rate": 2.6501162848634023e-06,
+      "loss": 0.5952,
+      "step": 2620
+    },
+    {
+      "epoch": 3.3,
+      "grad_norm": 0.9844443798065186,
+      "learning_rate": 2.6152217204549045e-06,
+      "loss": 0.5304,
+      "step": 2630
+    },
+    {
+      "epoch": 3.31,
+      "grad_norm": 1.0739080905914307,
+      "learning_rate": 2.580476868903197e-06,
+      "loss": 0.4941,
+      "step": 2640
+    },
+    {
+      "epoch": 3.32,
+      "grad_norm": 0.8901020884513855,
+      "learning_rate": 2.5458839114387146e-06,
+      "loss": 0.5418,
+      "step": 2650
+    },
+    {
+      "epoch": 3.34,
+      "grad_norm": 0.9322318434715271,
+      "learning_rate": 2.5114450197562214e-06,
+      "loss": 0.4952,
+      "step": 2660
+    },
+    {
+      "epoch": 3.35,
+      "grad_norm": 0.9946108460426331,
+      "learning_rate": 2.4771623558784474e-06,
+      "loss": 0.5155,
+      "step": 2670
+    },
+    {
+      "epoch": 3.36,
+      "grad_norm": 0.9886655211448669,
+      "learning_rate": 2.4430380720203785e-06,
+      "loss": 0.5374,
+      "step": 2680
+    },
+    {
+      "epoch": 3.38,
+      "grad_norm": 0.9973739385604858,
+      "learning_rate": 2.409074310454136e-06,
+      "loss": 0.5055,
+      "step": 2690
+    },
+    {
+      "epoch": 3.39,
+      "grad_norm": 0.9865736961364746,
+      "learning_rate": 2.375273203374489e-06,
+      "loss": 0.5287,
+      "step": 2700
+    },
+    {
+      "epoch": 3.4,
+      "grad_norm": 0.9992399215698242,
+      "learning_rate": 2.341636872764994e-06,
+      "loss": 0.5753,
+      "step": 2710
+    },
+    {
+      "epoch": 3.41,
+      "grad_norm": 1.0160070657730103,
+      "learning_rate": 2.3081674302647905e-06,
+      "loss": 0.5102,
+      "step": 2720
+    },
+    {
+      "epoch": 3.43,
+      "grad_norm": 0.7992890477180481,
+      "learning_rate": 2.2748669770360273e-06,
+      "loss": 0.5551,
+      "step": 2730
+    },
+    {
+      "epoch": 3.44,
+      "grad_norm": 1.0403869152069092,
+      "learning_rate": 2.2417376036319495e-06,
+      "loss": 0.5231,
+      "step": 2740
+    },
+    {
+      "epoch": 3.45,
+      "grad_norm": 0.9026238918304443,
+      "learning_rate": 2.2087813898656775e-06,
+      "loss": 0.5148,
+      "step": 2750
+    },
+    {
+      "epoch": 3.46,
+      "grad_norm": 1.078641414642334,
+      "learning_rate": 2.176000404679612e-06,
+      "loss": 0.5481,
+      "step": 2760
+    },
+    {
+      "epoch": 3.48,
+      "grad_norm": 0.8496494889259338,
+      "learning_rate": 2.1433967060155687e-06,
+      "loss": 0.4961,
+      "step": 2770
+    },
+    {
+      "epoch": 3.49,
+      "grad_norm": 0.9248107671737671,
+      "learning_rate": 2.110972340685577e-06,
+      "loss": 0.5402,
+      "step": 2780
+    },
+    {
+      "epoch": 3.5,
+      "grad_norm": 1.0704095363616943,
+      "learning_rate": 2.078729344243384e-06,
+      "loss": 0.5228,
+      "step": 2790
+    },
+    {
+      "epoch": 3.51,
+      "grad_norm": 0.8805019855499268,
+      "learning_rate": 2.0466697408566664e-06,
+      "loss": 0.545,
+      "step": 2800
+    },
+    {
+      "epoch": 3.51,
+      "eval_loss": 0.5704147815704346,
+      "eval_runtime": 77.346,
+      "eval_samples_per_second": 58.18,
+      "eval_steps_per_second": 1.823,
+      "step": 2800
+    },
+    {
+      "epoch": 3.53,
+      "grad_norm": 1.0108119249343872,
+      "learning_rate": 2.014795543179957e-06,
+      "loss": 0.5361,
+      "step": 2810
+    },
+    {
+      "epoch": 3.54,
+      "grad_norm": 1.01576828956604,
+      "learning_rate": 1.9831087522282865e-06,
+      "loss": 0.5552,
+      "step": 2820
+    },
+    {
+      "epoch": 3.55,
+      "grad_norm": 0.9220107197761536,
+      "learning_rate": 1.951611357251575e-06,
+      "loss": 0.5753,
+      "step": 2830
+    },
+    {
+      "epoch": 3.56,
+      "grad_norm": 0.9482864141464233,
+      "learning_rate": 1.92030533560974e-06,
+      "loss": 0.5339,
+      "step": 2840
+    },
+    {
+      "epoch": 3.58,
+      "grad_norm": 0.9235343933105469,
+      "learning_rate": 1.8891926526485643e-06,
+      "loss": 0.538,
+      "step": 2850
+    },
+    {
+      "epoch": 3.59,
+      "grad_norm": 0.6921684145927429,
+      "learning_rate": 1.8582752615763161e-06,
+      "loss": 0.5331,
+      "step": 2860
+    },
+    {
+      "epoch": 3.6,
+      "grad_norm": 0.930234968662262,
+      "learning_rate": 1.8275551033411209e-06,
+      "loss": 0.5226,
+      "step": 2870
+    },
+    {
+      "epoch": 3.61,
+      "grad_norm": 0.9894780516624451,
+      "learning_rate": 1.7970341065091246e-06,
+      "loss": 0.5294,
+      "step": 2880
+    },
+    {
+      "epoch": 3.63,
+      "grad_norm": 1.0227577686309814,
+      "learning_rate": 1.766714187143413e-06,
+      "loss": 0.5186,
+      "step": 2890
+    },
+    {
+      "epoch": 3.64,
+      "grad_norm": 0.9115012884140015,
+      "learning_rate": 1.736597248683724e-06,
+      "loss": 0.5156,
+      "step": 2900
+    },
+    {
+      "epoch": 3.65,
+      "grad_norm": 1.050728440284729,
+      "learning_rate": 1.7066851818269558e-06,
+      "loss": 0.521,
+      "step": 2910
+    },
+    {
+      "epoch": 3.66,
+      "grad_norm": 1.0954750776290894,
+      "learning_rate": 1.676979864408471e-06,
+      "loss": 0.5451,
+      "step": 2920
+    },
+    {
+      "epoch": 3.68,
+      "grad_norm": 0.7569468021392822,
+      "learning_rate": 1.647483161284203e-06,
+      "loss": 0.5765,
+      "step": 2930
+    },
+    {
+      "epoch": 3.69,
+      "grad_norm": 0.8553792238235474,
+      "learning_rate": 1.6181969242135926e-06,
+      "loss": 0.5335,
+      "step": 2940
+    },
+    {
+      "epoch": 3.7,
+      "grad_norm": 0.918138861656189,
+      "learning_rate": 1.5891229917433316e-06,
+      "loss": 0.5343,
+      "step": 2950
+    },
+    {
+      "epoch": 3.71,
+      "grad_norm": 0.988727867603302,
+      "learning_rate": 1.5602631890919428e-06,
+      "loss": 0.5288,
+      "step": 2960
+    },
+    {
+      "epoch": 3.73,
+      "grad_norm": 1.130867600440979,
+      "learning_rate": 1.531619328035197e-06,
+      "loss": 0.4952,
+      "step": 2970
+    },
+    {
+      "epoch": 3.74,
+      "grad_norm": 1.1033731698989868,
+      "learning_rate": 1.503193206792366e-06,
+      "loss": 0.5053,
+      "step": 2980
+    },
+    {
+      "epoch": 3.75,
+      "grad_norm": 1.0361392498016357,
+      "learning_rate": 1.4749866099133409e-06,
+      "loss": 0.555,
+      "step": 2990
+    },
+    {
+      "epoch": 3.76,
+      "grad_norm": 1.1811940670013428,
+      "learning_rate": 1.447001308166599e-06,
+      "loss": 0.5253,
+      "step": 3000
+    },
+    {
+      "epoch": 3.76,
+      "eval_loss": 0.5704381465911865,
+      "eval_runtime": 72.5701,
+      "eval_samples_per_second": 62.009,
+      "eval_steps_per_second": 1.943,
+      "step": 3000
+    },
+    {
+      "epoch": 3.78,
+      "grad_norm": 1.0649921894073486,
+      "learning_rate": 1.4192390584280347e-06,
+      "loss": 0.5397,
+      "step": 3010
+    },
+    {
+      "epoch": 3.79,
+      "grad_norm": 1.006032943725586,
+      "learning_rate": 1.3917016035706632e-06,
+      "loss": 0.5321,
+      "step": 3020
+    },
+    {
+      "epoch": 3.8,
+      "grad_norm": 1.1129940748214722,
+      "learning_rate": 1.3643906723552186e-06,
+      "loss": 0.5246,
+      "step": 3030
+    },
+    {
+      "epoch": 3.81,
+      "grad_norm": 1.1166002750396729,
+      "learning_rate": 1.3373079793216037e-06,
+      "loss": 0.5405,
+      "step": 3040
+    },
+    {
+      "epoch": 3.83,
+      "grad_norm": 0.9260038137435913,
+      "learning_rate": 1.3104552246812768e-06,
+      "loss": 0.543,
+      "step": 3050
+    },
+    {
+      "epoch": 3.84,
+      "grad_norm": 1.0943797826766968,
+      "learning_rate": 1.2838340942104944e-06,
+      "loss": 0.5147,
+      "step": 3060
+    },
+    {
+      "epoch": 3.85,
+      "grad_norm": 1.0058125257492065,
+      "learning_rate": 1.257446259144494e-06,
+      "loss": 0.4995,
+      "step": 3070
+    },
+    {
+      "epoch": 3.86,
+      "grad_norm": 1.022620677947998,
+      "learning_rate": 1.2312933760725753e-06,
+      "loss": 0.5485,
+      "step": 3080
+    },
+    {
+      "epoch": 3.88,
+      "grad_norm": 1.077143907546997,
+      "learning_rate": 1.2053770868340914e-06,
+      "loss": 0.5483,
+      "step": 3090
+    },
+    {
+      "epoch": 3.89,
+      "grad_norm": 1.3038913011550903,
+      "learning_rate": 1.17969901841539e-06,
+      "loss": 0.5388,
+      "step": 3100
+    },
+    {
+      "epoch": 3.9,
+      "grad_norm": 0.8922839760780334,
+      "learning_rate": 1.1542607828476665e-06,
+      "loss": 0.5404,
+      "step": 3110
+    },
+    {
+      "epoch": 3.91,
+      "grad_norm": 0.9709892272949219,
+      "learning_rate": 1.1290639771057648e-06,
+      "loss": 0.5204,
+      "step": 3120
+    },
+    {
+      "epoch": 3.93,
+      "grad_norm": 1.1010524034500122,
+      "learning_rate": 1.1041101830079159e-06,
+      "loss": 0.5429,
+      "step": 3130
+    },
+    {
+      "epoch": 3.94,
+      "grad_norm": 1.1041542291641235,
+      "learning_rate": 1.0794009671164484e-06,
+      "loss": 0.5256,
+      "step": 3140
+    },
+    {
+      "epoch": 3.95,
+      "grad_norm": 1.1114507913589478,
+      "learning_rate": 1.054937880639424e-06,
+      "loss": 0.5185,
+      "step": 3150
+    },
+    {
+      "epoch": 3.96,
+      "grad_norm": 0.9926356077194214,
+      "learning_rate": 1.030722459333267e-06,
+      "loss": 0.5161,
+      "step": 3160
+    },
+    {
+      "epoch": 3.98,
+      "grad_norm": 1.00425386428833,
+      "learning_rate": 1.0067562234063477e-06,
+      "loss": 0.5222,
+      "step": 3170
+    },
+    {
+      "epoch": 3.99,
+      "grad_norm": 1.0224225521087646,
+      "learning_rate": 9.830406774235474e-07,
+      "loss": 0.521,
+      "step": 3180
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 0.9822995662689209,
+      "learning_rate": 9.595773102118022e-07,
+      "loss": 0.528,
+      "step": 3190
+    },
+    {
+      "epoch": 4.02,
+      "grad_norm": 1.1153786182403564,
+      "learning_rate": 9.363675947666362e-07,
+      "loss": 0.5418,
+      "step": 3200
+    },
+    {
+      "epoch": 4.02,
+      "eval_loss": 0.5702469944953918,
+      "eval_runtime": 71.9125,
+      "eval_samples_per_second": 62.576,
+      "eval_steps_per_second": 1.961,
+      "step": 3200
+    },
+    {
+      "epoch": 4.03,
+      "grad_norm": 1.056578278541565,
+      "learning_rate": 9.134129881596926e-07,
+      "loss": 0.5326,
+      "step": 3210
+    },
+    {
+      "epoch": 4.04,
+      "grad_norm": 0.9958868026733398,
+      "learning_rate": 8.907149314472579e-07,
+      "loss": 0.558,
+      "step": 3220
+    },
+    {
+      "epoch": 4.05,
+      "grad_norm": 1.1032071113586426,
+      "learning_rate": 8.682748495797965e-07,
+      "loss": 0.5423,
+      "step": 3230
+    },
+    {
+      "epoch": 4.07,
+      "grad_norm": 1.0517561435699463,
+      "learning_rate": 8.460941513124876e-07,
+      "loss": 0.5189,
+      "step": 3240
+    },
+    {
+      "epoch": 4.08,
+      "grad_norm": 1.1289842128753662,
+      "learning_rate": 8.241742291168015e-07,
+      "loss": 0.5002,
+      "step": 3250
+    },
+    {
+      "epoch": 4.09,
+      "grad_norm": 0.9873390793800354,
+      "learning_rate": 8.025164590930623e-07,
+      "loss": 0.5589,
+      "step": 3260
+    },
+    {
+      "epoch": 4.1,
+      "grad_norm": 0.885833740234375,
+      "learning_rate": 7.811222008840719e-07,
+      "loss": 0.5031,
+      "step": 3270
+    },
+    {
+      "epoch": 4.12,
+      "grad_norm": 0.7554776072502136,
+      "learning_rate": 7.599927975897514e-07,
+      "loss": 0.5128,
+      "step": 3280
+    },
+    {
+      "epoch": 4.13,
+      "grad_norm": 1.0165417194366455,
+      "learning_rate": 7.391295756828187e-07,
+      "loss": 0.5369,
+      "step": 3290
+    },
+    {
+      "epoch": 4.14,
+      "grad_norm": 0.9958201050758362,
+      "learning_rate": 7.185338449255197e-07,
+      "loss": 0.4948,
+      "step": 3300
+    },
+    {
+      "epoch": 4.15,
+      "grad_norm": 1.1157667636871338,
+      "learning_rate": 6.982068982874019e-07,
+      "loss": 0.5181,
+      "step": 3310
+    },
+    {
+      "epoch": 4.17,
+      "grad_norm": 1.0274876356124878,
+      "learning_rate": 6.781500118641377e-07,
+      "loss": 0.5172,
+      "step": 3320
+    },
+    {
+      "epoch": 4.18,
+      "grad_norm": 0.9985108375549316,
+      "learning_rate": 6.583644447974241e-07,
+      "loss": 0.5304,
+      "step": 3330
+    },
+    {
+      "epoch": 4.19,
+      "grad_norm": 1.0142953395843506,
+      "learning_rate": 6.388514391959283e-07,
+      "loss": 0.5278,
+      "step": 3340
+    },
+    {
+      "epoch": 4.2,
+      "grad_norm": 1.0171020030975342,
+      "learning_rate": 6.196122200573073e-07,
+      "loss": 0.5269,
+      "step": 3350
+    },
+    {
+      "epoch": 4.22,
+      "grad_norm": 1.2027987241744995,
+      "learning_rate": 6.006479951913168e-07,
+      "loss": 0.5379,
+      "step": 3360
+    },
+    {
+      "epoch": 4.23,
+      "grad_norm": 1.064353108406067,
+      "learning_rate": 5.819599551439703e-07,
+      "loss": 0.4879,
+      "step": 3370
+    },
+    {
+      "epoch": 4.24,
+      "grad_norm": 1.0336158275604248,
+      "learning_rate": 5.635492731228098e-07,
+      "loss": 0.5363,
+      "step": 3380
+    },
+    {
+      "epoch": 4.25,
+      "grad_norm": 1.143526554107666,
+      "learning_rate": 5.45417104923251e-07,
+      "loss": 0.5371,
+      "step": 3390
+    },
+    {
+      "epoch": 4.27,
+      "grad_norm": 0.9199160933494568,
+      "learning_rate": 5.275645888560233e-07,
+      "loss": 0.5289,
+      "step": 3400
+    },
+    {
+      "epoch": 4.27,
+      "eval_loss": 0.570408046245575,
+      "eval_runtime": 72.3442,
+      "eval_samples_per_second": 62.203,
+      "eval_steps_per_second": 1.949,
+      "step": 3400
+    },
+    {
+      "epoch": 4.28,
+      "grad_norm": 1.0219011306762695,
+      "learning_rate": 5.099928456757036e-07,
+      "loss": 0.5115,
+      "step": 3410
+    },
+    {
+      "epoch": 4.29,
+      "grad_norm": 1.0187703371047974,
+      "learning_rate": 4.927029785103715e-07,
+      "loss": 0.5171,
+      "step": 3420
+    },
+    {
+      "epoch": 4.3,
+      "grad_norm": 1.006913423538208,
+      "learning_rate": 4.756960727923371e-07,
+      "loss": 0.5113,
+      "step": 3430
+    },
+    {
+      "epoch": 4.32,
+      "grad_norm": 1.0249481201171875,
+      "learning_rate": 4.5897319619001404e-07,
+      "loss": 0.552,
+      "step": 3440
+    },
+    {
+      "epoch": 4.33,
+      "grad_norm": 1.016176462173462,
+      "learning_rate": 4.42535398540887e-07,
+      "loss": 0.5246,
+      "step": 3450
+    },
+    {
+      "epoch": 4.34,
+      "grad_norm": 0.9975274205207825,
+      "learning_rate": 4.263837117856046e-07,
+      "loss": 0.5328,
+      "step": 3460
+    },
+    {
+      "epoch": 4.35,
+      "grad_norm": 0.8603295683860779,
+      "learning_rate": 4.10519149903198e-07,
+      "loss": 0.5409,
+      "step": 3470
+    },
+    {
+      "epoch": 4.37,
+      "grad_norm": 1.079770565032959,
+      "learning_rate": 3.949427088474189e-07,
+      "loss": 0.5506,
+      "step": 3480
+    },
+    {
+      "epoch": 4.38,
+      "grad_norm": 0.9214140176773071,
+      "learning_rate": 3.79655366484224e-07,
+      "loss": 0.5175,
+      "step": 3490
+    },
+    {
+      "epoch": 4.39,
+      "grad_norm": 0.9992761015892029,
+      "learning_rate": 3.646580825303786e-07,
+      "loss": 0.5197,
+      "step": 3500
+    },
+    {
+      "epoch": 4.4,
+      "grad_norm": 0.9221498370170593,
+      "learning_rate": 3.49951798493211e-07,
+      "loss": 0.5083,
+      "step": 3510
+    },
+    {
+      "epoch": 4.42,
+      "grad_norm": 1.0544548034667969,
+      "learning_rate": 3.355374376115017e-07,
+      "loss": 0.5187,
+      "step": 3520
+    },
+    {
+      "epoch": 4.43,
+      "grad_norm": 0.8989390134811401,
+      "learning_rate": 3.214159047975324e-07,
+      "loss": 0.5137,
+      "step": 3530
+    },
+    {
+      "epoch": 4.44,
+      "grad_norm": 0.9407683610916138,
+      "learning_rate": 3.0758808658026496e-07,
+      "loss": 0.5045,
+      "step": 3540
+    },
+    {
+      "epoch": 4.45,
+      "grad_norm": 1.034885287284851,
+      "learning_rate": 2.940548510496977e-07,
+      "loss": 0.5369,
+      "step": 3550
+    },
+    {
+      "epoch": 4.47,
+      "grad_norm": 0.9151519536972046,
+      "learning_rate": 2.8081704780236083e-07,
+      "loss": 0.5619,
+      "step": 3560
+    },
+    {
+      "epoch": 4.48,
+      "grad_norm": 1.22360098361969,
+      "learning_rate": 2.6787550788798243e-07,
+      "loss": 0.5078,
+      "step": 3570
+    },
+    {
+      "epoch": 4.49,
+      "grad_norm": 1.11081063747406,
+      "learning_rate": 2.55231043757318e-07,
+      "loss": 0.5391,
+      "step": 3580
+    },
+    {
+      "epoch": 4.5,
+      "grad_norm": 0.9431936144828796,
+      "learning_rate": 2.428844492111415e-07,
+      "loss": 0.581,
+      "step": 3590
+    },
+    {
+      "epoch": 4.52,
+      "grad_norm": 0.993818461894989,
+      "learning_rate": 2.3083649935041642e-07,
+      "loss": 0.4983,
+      "step": 3600
+    },
+    {
+      "epoch": 4.52,
+      "eval_loss": 0.570551335811615,
+      "eval_runtime": 73.4513,
+      "eval_samples_per_second": 61.265,
+      "eval_steps_per_second": 1.92,
+      "step": 3600
+    },
+    {
+      "epoch": 4.53,
+      "grad_norm": 0.9568440318107605,
+      "learning_rate": 2.1908795052763388e-07,
+      "loss": 0.5094,
+      "step": 3610
+    },
+    {
+      "epoch": 4.54,
+      "grad_norm": 1.0978604555130005,
+      "learning_rate": 2.0763954029933085e-07,
+      "loss": 0.5262,
+      "step": 3620
+    },
+    {
+      "epoch": 4.55,
+      "grad_norm": 0.9831346273422241,
+      "learning_rate": 1.9649198737978325e-07,
+      "loss": 0.5381,
+      "step": 3630
+    },
+    {
+      "epoch": 4.57,
+      "grad_norm": 1.1549886465072632,
+      "learning_rate": 1.856459915958958e-07,
+      "loss": 0.5069,
+      "step": 3640
+    },
+    {
+      "epoch": 4.58,
+      "grad_norm": 1.122920274734497,
+      "learning_rate": 1.7510223384325674e-07,
+      "loss": 0.5323,
+      "step": 3650
+    },
+    {
+      "epoch": 4.59,
+      "grad_norm": 1.0226640701293945,
+      "learning_rate": 1.6486137604339813e-07,
+      "loss": 0.5206,
+      "step": 3660
+    },
+    {
+      "epoch": 4.6,
+      "grad_norm": 0.9887501001358032,
+      "learning_rate": 1.5492406110224289e-07,
+      "loss": 0.549,
+      "step": 3670
+    },
+    {
+      "epoch": 4.62,
+      "grad_norm": 1.069175124168396,
+      "learning_rate": 1.4529091286973994e-07,
+      "loss": 0.535,
+      "step": 3680
+    },
+    {
+      "epoch": 4.63,
+      "grad_norm": 1.0605347156524658,
+      "learning_rate": 1.3596253610070165e-07,
+      "loss": 0.4874,
+      "step": 3690
+    },
+    {
+      "epoch": 4.64,
+      "grad_norm": 1.1193597316741943,
+      "learning_rate": 1.2693951641683856e-07,
+      "loss": 0.5137,
+      "step": 3700
+    },
+    {
+      "epoch": 4.65,
+      "grad_norm": 1.0010859966278076,
+      "learning_rate": 1.1822242026999443e-07,
+      "loss": 0.4834,
+      "step": 3710
+    },
+    {
+      "epoch": 4.67,
+      "grad_norm": 1.1622557640075684,
+      "learning_rate": 1.0981179490658522e-07,
+      "loss": 0.5739,
+      "step": 3720
+    },
+    {
+      "epoch": 4.68,
+      "grad_norm": 1.0148837566375732,
+      "learning_rate": 1.0170816833324549e-07,
+      "loss": 0.5182,
+      "step": 3730
+    },
+    {
+      "epoch": 4.69,
+      "grad_norm": 1.1132534742355347,
+      "learning_rate": 9.391204928367648e-08,
+      "loss": 0.517,
+      "step": 3740
+    },
+    {
+      "epoch": 4.71,
+      "grad_norm": 1.0367366075515747,
+      "learning_rate": 8.642392718671455e-08,
+      "loss": 0.5191,
+      "step": 3750
+    },
+    {
+      "epoch": 4.72,
+      "grad_norm": 1.0188932418823242,
+      "learning_rate": 7.92442721355996e-08,
+      "loss": 0.5277,
+      "step": 3760
+    },
+    {
+      "epoch": 4.73,
+      "grad_norm": 1.0666061639785767,
+      "learning_rate": 7.237353485846865e-08,
+      "loss": 0.5272,
+      "step": 3770
+    },
+    {
+      "epoch": 4.74,
+      "grad_norm": 0.9185099601745605,
+      "learning_rate": 6.58121466900552e-08,
+      "loss": 0.5114,
+      "step": 3780
+    },
+    {
+      "epoch": 4.76,
+      "grad_norm": 1.1004579067230225,
+      "learning_rate": 5.9560519544614725e-08,
+      "loss": 0.5088,
+      "step": 3790
+    },
+    {
+      "epoch": 4.77,
+      "grad_norm": 1.0939706563949585,
+      "learning_rate": 5.36190458900615e-08,
+      "loss": 0.5371,
+      "step": 3800
+    },
+    {
+      "epoch": 4.77,
+      "eval_loss": 0.5705362558364868,
+      "eval_runtime": 72.0164,
+      "eval_samples_per_second": 62.486,
+      "eval_steps_per_second": 1.958,
+      "step": 3800
+    },
+    {
+      "epoch": 4.78,
+      "grad_norm": 1.0084924697875977,
+      "learning_rate": 4.798809872333276e-08,
+      "loss": 0.5561,
+      "step": 3810
+    },
+    {
+      "epoch": 4.79,
+      "grad_norm": 1.118712067604065,
+      "learning_rate": 4.266803154697241e-08,
+      "loss": 0.5511,
+      "step": 3820
+    },
+    {
+      "epoch": 4.81,
+      "grad_norm": 1.0696793794631958,
+      "learning_rate": 3.765917834693711e-08,
+      "loss": 0.4753,
+      "step": 3830
+    },
+    {
+      "epoch": 4.82,
+      "grad_norm": 0.9693195223808289,
+      "learning_rate": 3.296185357162973e-08,
+      "loss": 0.5308,
+      "step": 3840
+    },
+    {
+      "epoch": 4.83,
+      "grad_norm": 1.145647406578064,
+      "learning_rate": 2.8576352112158478e-08,
+      "loss": 0.5174,
+      "step": 3850
+    },
+    {
+      "epoch": 4.84,
+      "grad_norm": 0.9660029411315918,
+      "learning_rate": 2.4502949283826683e-08,
+      "loss": 0.4866,
+      "step": 3860
+    },
+    {
+      "epoch": 4.86,
+      "grad_norm": 0.9987422823905945,
+      "learning_rate": 2.074190080884386e-08,
+      "loss": 0.5169,
+      "step": 3870
+    },
+    {
+      "epoch": 4.87,
+      "grad_norm": 0.9800373911857605,
+      "learning_rate": 1.729344280027745e-08,
+      "loss": 0.5065,
+      "step": 3880
+    },
+    {
+      "epoch": 4.88,
+      "grad_norm": 1.1359976530075073,
+      "learning_rate": 1.4157791747225202e-08,
+      "loss": 0.5496,
+      "step": 3890
+    },
+    {
+      "epoch": 4.89,
+      "grad_norm": 0.9586722254753113,
+      "learning_rate": 1.133514450122719e-08,
+      "loss": 0.5138,
+      "step": 3900
+    },
+    {
+      "epoch": 4.91,
+      "grad_norm": 1.1620326042175293,
+      "learning_rate": 8.825678263906789e-09,
+      "loss": 0.5573,
+      "step": 3910
+    },
+    {
+      "epoch": 4.92,
+      "grad_norm": 0.8677152395248413,
+      "learning_rate": 6.629550575847355e-09,
+      "loss": 0.4891,
+      "step": 3920
+    },
+    {
+      "epoch": 4.93,
+      "grad_norm": 1.0054067373275757,
+      "learning_rate": 4.746899306699582e-09,
+      "loss": 0.5325,
+      "step": 3930
+    },
+    {
+      "epoch": 4.94,
+      "grad_norm": 1.2144650220870972,
+      "learning_rate": 3.1778426465289838e-09,
+      "loss": 0.4983,
+      "step": 3940
+    },
+    {
+      "epoch": 4.96,
+      "grad_norm": 1.1614487171173096,
+      "learning_rate": 1.9224790983940477e-09,
+      "loss": 0.5108,
+      "step": 3950
+    },
+    {
+      "epoch": 4.97,
+      "grad_norm": 0.9379814267158508,
+      "learning_rate": 9.808874721628503e-10,
+      "loss": 0.5122,
+      "step": 3960
+    },
+    {
+      "epoch": 4.98,
+      "grad_norm": 1.0148200988769531,
+      "learning_rate": 3.5312687956756684e-10,
+      "loss": 0.5337,
+      "step": 3970
+    },
+    {
+      "epoch": 4.99,
+      "grad_norm": 1.089341640472412,
+      "learning_rate": 3.923673048966681e-11,
+      "loss": 0.529,
+      "step": 3980
+    },
+    {
+      "epoch": 5.0,
+      "step": 3985,
+      "total_flos": 2.2824671149817856e+18,
+      "train_loss": 0.569244503227052,
+      "train_runtime": 8741.9073,
+      "train_samples_per_second": 14.585,
+      "train_steps_per_second": 0.456
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 3985,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 5,
+  "save_steps": 1000,
+  "total_flos": 2.2824671149817856e+18,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}

llama2_13b_peft/news_commentary_es/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f945155f44888fffb9611998b5a2cb53a0ecf7e194d1ec8ea3f4941fc7e5b93
+size 5112

llama2_13b_peft/object_counting/README.md

@@ -0,0 +1,68 @@
+---
+license: other
+library_name: peft
+tags:
+- llama-factory
+- lora
+- generated_from_trainer
+base_model: /data1/model/llama2/meta-llama/Llama2-13b
+model-index:
+- name: object_counting_no_sys
+  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. -->
+
+# object_counting_no_sys
+
+This model is a fine-tuned version of [/data1/model/llama2/meta-llama/Llama2-13b](https://huggingface.co//data1/model/llama2/meta-llama/Llama2-13b) on the object_counting_no_sys dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.2107
+
+## 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: 0.0001
+- train_batch_size: 4
+- eval_batch_size: 4
+- seed: 42
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: cosine
+- lr_scheduler_warmup_steps: 20
+- num_epochs: 5.0
+
+### Training results
+
+| Training Loss | Epoch  | Step | Validation Loss |
+|:-------------:|:------:|:----:|:---------------:|
+| 0.477         | 0.5263 | 100  | 0.4622          |
+| 0.4647        | 1.0526 | 200  | 0.2513          |
+| 0.5231        | 1.5789 | 300  | 0.3432          |
+| 0.4381        | 2.1053 | 400  | 0.2107          |
+| 0.2159        | 2.6316 | 500  | 0.3666          |
+| 0.1934        | 3.1579 | 600  | 0.2625          |
+| 0.2701        | 3.6842 | 700  | 0.2854          |
+
+
+### Framework versions
+
+- PEFT 0.10.0
+- Transformers 4.40.0
+- Pytorch 2.2.1
+- Datasets 2.18.0
+- Tokenizers 0.19.1

llama2_13b_peft/object_counting/adapter_config.json

@@ -0,0 +1,34 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "/data1/model/llama2/meta-llama/Llama2-13b",
+  "bias": "none",
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_dropout": 0.0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "gate_proj",
+    "v_proj",
+    "q_proj",
+    "up_proj",
+    "o_proj",
+    "k_proj",
+    "down_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

llama2_13b_peft/object_counting/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6ec85f92f9090e17b7da8fd52d43bbdc1cfa184a3c8bd57e86f55941c22c564d
+size 125248064

llama2_13b_peft/object_counting/all_results.json

@@ -0,0 +1,12 @@
+{
+    "epoch": 3.6842105263157894,
+    "eval_loss": 0.21073441207408905,
+    "eval_runtime": 0.6962,
+    "eval_samples_per_second": 57.458,
+    "eval_steps_per_second": 14.365,
+    "total_flos": 1.461153753759744e+16,
+    "train_loss": 0.5859067852156503,
+    "train_runtime": 192.1858,
+    "train_samples_per_second": 19.773,
+    "train_steps_per_second": 4.943
+}

llama2_13b_peft/object_counting/eval_results.json

@@ -0,0 +1,7 @@
+{
+    "epoch": 3.6842105263157894,
+    "eval_loss": 0.21073441207408905,
+    "eval_runtime": 0.6962,
+    "eval_samples_per_second": 57.458,
+    "eval_steps_per_second": 14.365
+}

llama2_13b_peft/object_counting/special_tokens_map.json

@@ -0,0 +1,24 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "</s>",
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

llama2_13b_peft/object_counting/tokenizer.model

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

llama2_13b_peft/object_counting/tokenizer_config.json

@@ -0,0 +1,45 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "add_prefix_space": true,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "chat_template": "{% if messages[0]['role'] == 'system' %}{% set system_message = messages[0]['content'] %}{% endif %}{% if system_message is defined %}{{ system_message }}{% endif %}{% for message in messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ content }}{% elif message['role'] == 'assistant' %}{{ content + '\\n' }}{% endif %}{% endfor %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "</s>",
+  "legacy": true,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "</s>",
+  "padding_side": "right",
+  "sp_model_kwargs": {},
+  "spaces_between_special_tokens": false,
+  "split_special_tokens": false,
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": false
+}

llama2_13b_peft/object_counting/train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 3.6842105263157894,
+    "total_flos": 1.461153753759744e+16,
+    "train_loss": 0.5859067852156503,
+    "train_runtime": 192.1858,
+    "train_samples_per_second": 19.773,
+    "train_steps_per_second": 4.943
+}

llama2_13b_peft/object_counting/trainer_log.jsonl

@@ -0,0 +1,79 @@
+{"current_steps": 10, "total_steps": 950, "loss": 10.645, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5e-05, "epoch": 0.05263157894736842, "percentage": 1.05, "elapsed_time": "0:00:03", "remaining_time": "0:06:06"}
+{"current_steps": 20, "total_steps": 950, "loss": 4.9002, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 0.0001, "epoch": 0.10526315789473684, "percentage": 2.11, "elapsed_time": "0:00:06", "remaining_time": "0:04:58"}
+{"current_steps": 30, "total_steps": 950, "loss": 0.9012, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.99714745464859e-05, "epoch": 0.15789473684210525, "percentage": 3.16, "elapsed_time": "0:00:09", "remaining_time": "0:04:36"}
+{"current_steps": 40, "total_steps": 950, "loss": 0.7022, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.988593073400354e-05, "epoch": 0.21052631578947367, "percentage": 4.21, "elapsed_time": "0:00:11", "remaining_time": "0:04:22"}
+{"current_steps": 50, "total_steps": 950, "loss": 0.7425, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.974346616959476e-05, "epoch": 0.2631578947368421, "percentage": 5.26, "elapsed_time": "0:00:14", "remaining_time": "0:04:14"}
+{"current_steps": 60, "total_steps": 950, "loss": 0.7609, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.954424340791196e-05, "epoch": 0.3157894736842105, "percentage": 6.32, "elapsed_time": "0:00:16", "remaining_time": "0:04:07"}
+{"current_steps": 70, "total_steps": 950, "loss": 0.6316, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.928848976574019e-05, "epoch": 0.3684210526315789, "percentage": 7.37, "elapsed_time": "0:00:19", "remaining_time": "0:04:01"}
+{"current_steps": 80, "total_steps": 950, "loss": 0.6406, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.897649706262473e-05, "epoch": 0.42105263157894735, "percentage": 8.42, "elapsed_time": "0:00:21", "remaining_time": "0:03:56"}
+{"current_steps": 90, "total_steps": 950, "loss": 0.5597, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.860862128789953e-05, "epoch": 0.47368421052631576, "percentage": 9.47, "elapsed_time": "0:00:24", "remaining_time": "0:03:52"}
+{"current_steps": 100, "total_steps": 950, "loss": 0.477, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.818528219449705e-05, "epoch": 0.5263157894736842, "percentage": 10.53, "elapsed_time": "0:00:26", "remaining_time": "0:03:48"}
+{"current_steps": 100, "total_steps": 950, "loss": null, "eval_loss": 0.46219402551651, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 0.5263157894736842, "percentage": 10.53, "elapsed_time": "0:00:26", "remaining_time": "0:03:48"}
+{"current_steps": 110, "total_steps": 950, "loss": 0.4798, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.770696282000244e-05, "epoch": 0.5789473684210527, "percentage": 11.58, "elapsed_time": "0:00:30", "remaining_time": "0:03:53"}
+{"current_steps": 120, "total_steps": 950, "loss": 0.5097, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.717420893549902e-05, "epoch": 0.631578947368421, "percentage": 12.63, "elapsed_time": "0:00:33", "remaining_time": "0:03:49"}
+{"current_steps": 130, "total_steps": 950, "loss": 0.6507, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.658762842283343e-05, "epoch": 0.6842105263157895, "percentage": 13.68, "elapsed_time": "0:00:35", "remaining_time": "0:03:45"}
+{"current_steps": 140, "total_steps": 950, "loss": 0.4961, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.594789058101153e-05, "epoch": 0.7368421052631579, "percentage": 14.74, "elapsed_time": "0:00:38", "remaining_time": "0:03:41"}
+{"current_steps": 150, "total_steps": 950, "loss": 0.3643, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.525572536251607e-05, "epoch": 0.7894736842105263, "percentage": 15.79, "elapsed_time": "0:00:40", "remaining_time": "0:03:37"}
+{"current_steps": 160, "total_steps": 950, "loss": 0.5859, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.451192254041758e-05, "epoch": 0.8421052631578947, "percentage": 16.84, "elapsed_time": "0:00:43", "remaining_time": "0:03:33"}
+{"current_steps": 170, "total_steps": 950, "loss": 0.4765, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.371733080722911e-05, "epoch": 0.8947368421052632, "percentage": 17.89, "elapsed_time": "0:00:45", "remaining_time": "0:03:30"}
+{"current_steps": 180, "total_steps": 950, "loss": 0.4496, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.287285680653254e-05, "epoch": 0.9473684210526315, "percentage": 18.95, "elapsed_time": "0:00:48", "remaining_time": "0:03:27"}
+{"current_steps": 190, "total_steps": 950, "loss": 0.42, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.197946409848194e-05, "epoch": 1.0, "percentage": 20.0, "elapsed_time": "0:00:50", "remaining_time": "0:03:23"}
+{"current_steps": 200, "total_steps": 950, "loss": 0.4647, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.103817206036382e-05, "epoch": 1.0526315789473684, "percentage": 21.05, "elapsed_time": "0:00:53", "remaining_time": "0:03:20"}
+{"current_steps": 200, "total_steps": 950, "loss": null, "eval_loss": 0.2513052821159363, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 1.0526315789473684, "percentage": 21.05, "elapsed_time": "0:00:53", "remaining_time": "0:03:20"}
+{"current_steps": 210, "total_steps": 950, "loss": 0.3552, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 9.005005472346924e-05, "epoch": 1.1052631578947367, "percentage": 22.11, "elapsed_time": "0:00:57", "remaining_time": "0:03:21"}
+{"current_steps": 220, "total_steps": 950, "loss": 0.5985, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 8.90162395476046e-05, "epoch": 1.1578947368421053, "percentage": 23.16, "elapsed_time": "0:00:59", "remaining_time": "0:03:18"}
+{"current_steps": 230, "total_steps": 950, "loss": 0.6665, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 8.793790613463955e-05, "epoch": 1.2105263157894737, "percentage": 24.21, "elapsed_time": "0:01:02", "remaining_time": "0:03:15"}
+{"current_steps": 240, "total_steps": 950, "loss": 0.4249, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 8.681628488255986e-05, "epoch": 1.263157894736842, "percentage": 25.26, "elapsed_time": "0:01:04", "remaining_time": "0:03:11"}
+{"current_steps": 250, "total_steps": 950, "loss": 0.4735, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 8.565265558156101e-05, "epoch": 1.3157894736842106, "percentage": 26.32, "elapsed_time": "0:01:07", "remaining_time": "0:03:08"}
+{"current_steps": 260, "total_steps": 950, "loss": 0.265, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 8.444834595378434e-05, "epoch": 1.368421052631579, "percentage": 27.37, "elapsed_time": "0:01:09", "remaining_time": "0:03:05"}
+{"current_steps": 270, "total_steps": 950, "loss": 0.5925, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 8.320473013836196e-05, "epoch": 1.4210526315789473, "percentage": 28.42, "elapsed_time": "0:01:12", "remaining_time": "0:03:02"}
+{"current_steps": 280, "total_steps": 950, "loss": 0.4528, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 8.192322712349917e-05, "epoch": 1.4736842105263157, "percentage": 29.47, "elapsed_time": "0:01:14", "remaining_time": "0:02:59"}
+{"current_steps": 290, "total_steps": 950, "loss": 0.3558, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 8.060529912738315e-05, "epoch": 1.526315789473684, "percentage": 30.53, "elapsed_time": "0:01:17", "remaining_time": "0:02:56"}
+{"current_steps": 300, "total_steps": 950, "loss": 0.5231, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.925244992976538e-05, "epoch": 1.5789473684210527, "percentage": 31.58, "elapsed_time": "0:01:20", "remaining_time": "0:02:53"}
+{"current_steps": 300, "total_steps": 950, "loss": null, "eval_loss": 0.34315773844718933, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 1.5789473684210527, "percentage": 31.58, "elapsed_time": "0:01:20", "remaining_time": "0:02:53"}
+{"current_steps": 310, "total_steps": 950, "loss": 0.3751, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.786622315612183e-05, "epoch": 1.631578947368421, "percentage": 32.63, "elapsed_time": "0:01:23", "remaining_time": "0:02:52"}
+{"current_steps": 320, "total_steps": 950, "loss": 0.3844, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.644820051634812e-05, "epoch": 1.6842105263157894, "percentage": 33.68, "elapsed_time": "0:01:26", "remaining_time": "0:02:49"}
+{"current_steps": 330, "total_steps": 950, "loss": 0.5589, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.500000000000001e-05, "epoch": 1.736842105263158, "percentage": 34.74, "elapsed_time": "0:01:28", "remaining_time": "0:02:46"}
+{"current_steps": 340, "total_steps": 950, "loss": 0.2688, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.35232740301378e-05, "epoch": 1.7894736842105263, "percentage": 35.79, "elapsed_time": "0:01:31", "remaining_time": "0:02:43"}
+{"current_steps": 350, "total_steps": 950, "loss": 0.2528, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.201970757788172e-05, "epoch": 1.8421052631578947, "percentage": 36.84, "elapsed_time": "0:01:33", "remaining_time": "0:02:40"}
+{"current_steps": 360, "total_steps": 950, "loss": 0.2398, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 7.049101623982937e-05, "epoch": 1.8947368421052633, "percentage": 37.89, "elapsed_time": "0:01:36", "remaining_time": "0:02:38"}
+{"current_steps": 370, "total_steps": 950, "loss": 0.4979, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 6.89389442805288e-05, "epoch": 1.9473684210526314, "percentage": 38.95, "elapsed_time": "0:01:39", "remaining_time": "0:02:35"}
+{"current_steps": 380, "total_steps": 950, "loss": 0.5478, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 6.736526264224101e-05, "epoch": 2.0, "percentage": 40.0, "elapsed_time": "0:01:41", "remaining_time": "0:02:32"}
+{"current_steps": 390, "total_steps": 950, "loss": 0.4545, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 6.577176692426279e-05, "epoch": 2.0526315789473686, "percentage": 41.05, "elapsed_time": "0:01:44", "remaining_time": "0:02:29"}
+{"current_steps": 400, "total_steps": 950, "loss": 0.4381, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 6.416027533411519e-05, "epoch": 2.1052631578947367, "percentage": 42.11, "elapsed_time": "0:01:46", "remaining_time": "0:02:26"}
+{"current_steps": 400, "total_steps": 950, "loss": null, "eval_loss": 0.21073441207408905, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 2.1052631578947367, "percentage": 42.11, "elapsed_time": "0:01:46", "remaining_time": "0:02:26"}
+{"current_steps": 410, "total_steps": 950, "loss": 0.2746, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 6.253262661293604e-05, "epoch": 2.1578947368421053, "percentage": 43.16, "elapsed_time": "0:01:50", "remaining_time": "0:02:25"}
+{"current_steps": 420, "total_steps": 950, "loss": 0.2276, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 6.0890677937442574e-05, "epoch": 2.2105263157894735, "percentage": 44.21, "elapsed_time": "0:01:52", "remaining_time": "0:02:22"}
+{"current_steps": 430, "total_steps": 950, "loss": 0.2474, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5.923630280085948e-05, "epoch": 2.263157894736842, "percentage": 45.26, "elapsed_time": "0:01:55", "remaining_time": "0:02:19"}
+{"current_steps": 440, "total_steps": 950, "loss": 0.2644, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5.757138887522884e-05, "epoch": 2.3157894736842106, "percentage": 46.32, "elapsed_time": "0:01:57", "remaining_time": "0:02:16"}
+{"current_steps": 450, "total_steps": 950, "loss": 0.3294, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5.5897835857542317e-05, "epoch": 2.3684210526315788, "percentage": 47.37, "elapsed_time": "0:02:00", "remaining_time": "0:02:13"}
+{"current_steps": 460, "total_steps": 950, "loss": 0.5553, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5.4217553302152237e-05, "epoch": 2.4210526315789473, "percentage": 48.42, "elapsed_time": "0:02:02", "remaining_time": "0:02:10"}
+{"current_steps": 470, "total_steps": 950, "loss": 0.1107, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5.2532458441935636e-05, "epoch": 2.473684210526316, "percentage": 49.47, "elapsed_time": "0:02:05", "remaining_time": "0:02:08"}
+{"current_steps": 480, "total_steps": 950, "loss": 0.2507, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 5.084447400069655e-05, "epoch": 2.526315789473684, "percentage": 50.53, "elapsed_time": "0:02:08", "remaining_time": "0:02:05"}
+{"current_steps": 490, "total_steps": 950, "loss": 0.1839, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.915552599930345e-05, "epoch": 2.5789473684210527, "percentage": 51.58, "elapsed_time": "0:02:10", "remaining_time": "0:02:02"}
+{"current_steps": 500, "total_steps": 950, "loss": 0.2159, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.746754155806437e-05, "epoch": 2.6315789473684212, "percentage": 52.63, "elapsed_time": "0:02:13", "remaining_time": "0:01:59"}
+{"current_steps": 500, "total_steps": 950, "loss": null, "eval_loss": 0.36658158898353577, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 2.6315789473684212, "percentage": 52.63, "elapsed_time": "0:02:13", "remaining_time": "0:01:59"}
+{"current_steps": 510, "total_steps": 950, "loss": 0.1742, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.578244669784777e-05, "epoch": 2.6842105263157894, "percentage": 53.68, "elapsed_time": "0:02:16", "remaining_time": "0:01:58"}
+{"current_steps": 520, "total_steps": 950, "loss": 0.3915, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.410216414245771e-05, "epoch": 2.736842105263158, "percentage": 54.74, "elapsed_time": "0:02:19", "remaining_time": "0:01:55"}
+{"current_steps": 530, "total_steps": 950, "loss": 0.2055, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.2428611124771184e-05, "epoch": 2.7894736842105265, "percentage": 55.79, "elapsed_time": "0:02:22", "remaining_time": "0:01:52"}
+{"current_steps": 540, "total_steps": 950, "loss": 0.3395, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 4.076369719914055e-05, "epoch": 2.8421052631578947, "percentage": 56.84, "elapsed_time": "0:02:24", "remaining_time": "0:01:49"}
+{"current_steps": 550, "total_steps": 950, "loss": 0.2351, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.9109322062557424e-05, "epoch": 2.8947368421052633, "percentage": 57.89, "elapsed_time": "0:02:27", "remaining_time": "0:01:47"}
+{"current_steps": 560, "total_steps": 950, "loss": 0.1548, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.746737338706397e-05, "epoch": 2.9473684210526314, "percentage": 58.95, "elapsed_time": "0:02:29", "remaining_time": "0:01:44"}
+{"current_steps": 570, "total_steps": 950, "loss": 0.1734, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.58397246658848e-05, "epoch": 3.0, "percentage": 60.0, "elapsed_time": "0:02:32", "remaining_time": "0:01:41"}
+{"current_steps": 580, "total_steps": 950, "loss": 0.1015, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.422823307573722e-05, "epoch": 3.0526315789473686, "percentage": 61.05, "elapsed_time": "0:02:35", "remaining_time": "0:01:38"}
+{"current_steps": 590, "total_steps": 950, "loss": 0.1705, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.263473735775899e-05, "epoch": 3.1052631578947367, "percentage": 62.11, "elapsed_time": "0:02:37", "remaining_time": "0:01:36"}
+{"current_steps": 600, "total_steps": 950, "loss": 0.1934, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 3.10610557194712e-05, "epoch": 3.1578947368421053, "percentage": 63.16, "elapsed_time": "0:02:40", "remaining_time": "0:01:33"}
+{"current_steps": 600, "total_steps": 950, "loss": null, "eval_loss": 0.26252931356430054, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 3.1578947368421053, "percentage": 63.16, "elapsed_time": "0:02:40", "remaining_time": "0:01:33"}
+{"current_steps": 610, "total_steps": 950, "loss": 0.1024, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.950898376017064e-05, "epoch": 3.2105263157894735, "percentage": 64.21, "elapsed_time": "0:02:43", "remaining_time": "0:01:31"}
+{"current_steps": 620, "total_steps": 950, "loss": 0.1786, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.798029242211828e-05, "epoch": 3.263157894736842, "percentage": 65.26, "elapsed_time": "0:02:46", "remaining_time": "0:01:28"}
+{"current_steps": 630, "total_steps": 950, "loss": 0.1907, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.6476725969862227e-05, "epoch": 3.3157894736842106, "percentage": 66.32, "elapsed_time": "0:02:49", "remaining_time": "0:01:25"}
+{"current_steps": 640, "total_steps": 950, "loss": 0.125, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.500000000000001e-05, "epoch": 3.3684210526315788, "percentage": 67.37, "elapsed_time": "0:02:51", "remaining_time": "0:01:23"}
+{"current_steps": 650, "total_steps": 950, "loss": 0.1509, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.3551799483651894e-05, "epoch": 3.4210526315789473, "percentage": 68.42, "elapsed_time": "0:02:54", "remaining_time": "0:01:20"}
+{"current_steps": 660, "total_steps": 950, "loss": 0.1502, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.2133776843878186e-05, "epoch": 3.473684210526316, "percentage": 69.47, "elapsed_time": "0:02:56", "remaining_time": "0:01:17"}
+{"current_steps": 670, "total_steps": 950, "loss": 0.1458, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 2.074755007023461e-05, "epoch": 3.526315789473684, "percentage": 70.53, "elapsed_time": "0:02:59", "remaining_time": "0:01:14"}
+{"current_steps": 680, "total_steps": 950, "loss": 0.2181, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.9394700872616855e-05, "epoch": 3.5789473684210527, "percentage": 71.58, "elapsed_time": "0:03:01", "remaining_time": "0:01:12"}
+{"current_steps": 690, "total_steps": 950, "loss": 0.0979, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.807677287650083e-05, "epoch": 3.6315789473684212, "percentage": 72.63, "elapsed_time": "0:03:04", "remaining_time": "0:01:09"}
+{"current_steps": 700, "total_steps": 950, "loss": 0.2701, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": 1.6795269861638042e-05, "epoch": 3.6842105263157894, "percentage": 73.68, "elapsed_time": "0:03:06", "remaining_time": "0:01:06"}
+{"current_steps": 700, "total_steps": 950, "loss": null, "eval_loss": 0.2853814661502838, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 3.6842105263157894, "percentage": 73.68, "elapsed_time": "0:03:06", "remaining_time": "0:01:06"}
+{"current_steps": 700, "total_steps": 950, "loss": null, "eval_loss": null, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 3.6842105263157894, "percentage": 73.68, "elapsed_time": "0:03:06", "remaining_time": "0:01:06"}
+{"current_steps": 10, "total_steps": 10, "loss": null, "eval_loss": 0.21073441207408905, "predict_loss": null, "reward": null, "accuracy": null, "learning_rate": null, "epoch": 3.6842105263157894, "percentage": 100.0, "elapsed_time": "0:03:09", "remaining_time": "0:00:00"}

llama2_13b_peft/object_counting/trainer_state.json

@@ -0,0 +1,576 @@
+{
+  "best_metric": 0.21073441207408905,
+  "best_model_checkpoint": "ckpt/llama2_13b_fuze27_no_sys/object_counting_no_sys/checkpoint-400",
+  "epoch": 3.6842105263157894,
+  "eval_steps": 100,
+  "global_step": 700,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.05263157894736842,
+      "grad_norm": 6.26442289352417,
+      "learning_rate": 5e-05,
+      "loss": 10.645,
+      "step": 10
+    },
+    {
+      "epoch": 0.10526315789473684,
+      "grad_norm": 4.90908145904541,
+      "learning_rate": 0.0001,
+      "loss": 4.9002,
+      "step": 20
+    },
+    {
+      "epoch": 0.15789473684210525,
+      "grad_norm": 7.5457682609558105,
+      "learning_rate": 9.99714745464859e-05,
+      "loss": 0.9012,
+      "step": 30
+    },
+    {
+      "epoch": 0.21052631578947367,
+      "grad_norm": 9.139291763305664,
+      "learning_rate": 9.988593073400354e-05,
+      "loss": 0.7022,
+      "step": 40
+    },
+    {
+      "epoch": 0.2631578947368421,
+      "grad_norm": 5.00139856338501,
+      "learning_rate": 9.974346616959476e-05,
+      "loss": 0.7425,
+      "step": 50
+    },
+    {
+      "epoch": 0.3157894736842105,
+      "grad_norm": 6.002871990203857,
+      "learning_rate": 9.954424340791196e-05,
+      "loss": 0.7609,
+      "step": 60
+    },
+    {
+      "epoch": 0.3684210526315789,
+      "grad_norm": 6.101377010345459,
+      "learning_rate": 9.928848976574019e-05,
+      "loss": 0.6316,
+      "step": 70
+    },
+    {
+      "epoch": 0.42105263157894735,
+      "grad_norm": 5.7651824951171875,
+      "learning_rate": 9.897649706262473e-05,
+      "loss": 0.6406,
+      "step": 80
+    },
+    {
+      "epoch": 0.47368421052631576,
+      "grad_norm": 2.46134614944458,
+      "learning_rate": 9.860862128789953e-05,
+      "loss": 0.5597,
+      "step": 90
+    },
+    {
+      "epoch": 0.5263157894736842,
+      "grad_norm": 4.3927178382873535,
+      "learning_rate": 9.818528219449705e-05,
+      "loss": 0.477,
+      "step": 100
+    },
+    {
+      "epoch": 0.5263157894736842,
+      "eval_loss": 0.46219402551651,
+      "eval_runtime": 0.6936,
+      "eval_samples_per_second": 57.669,
+      "eval_steps_per_second": 14.417,
+      "step": 100
+    },
+    {
+      "epoch": 0.5789473684210527,
+      "grad_norm": 6.058330059051514,
+      "learning_rate": 9.770696282000244e-05,
+      "loss": 0.4798,
+      "step": 110
+    },
+    {
+      "epoch": 0.631578947368421,
+      "grad_norm": 4.410764217376709,
+      "learning_rate": 9.717420893549902e-05,
+      "loss": 0.5097,
+      "step": 120
+    },
+    {
+      "epoch": 0.6842105263157895,
+      "grad_norm": 1.7051620483398438,
+      "learning_rate": 9.658762842283343e-05,
+      "loss": 0.6507,
+      "step": 130
+    },
+    {
+      "epoch": 0.7368421052631579,
+      "grad_norm": 3.638519763946533,
+      "learning_rate": 9.594789058101153e-05,
+      "loss": 0.4961,
+      "step": 140
+    },
+    {
+      "epoch": 0.7894736842105263,
+      "grad_norm": 1.8080860376358032,
+      "learning_rate": 9.525572536251607e-05,
+      "loss": 0.3643,
+      "step": 150
+    },
+    {
+      "epoch": 0.8421052631578947,
+      "grad_norm": 4.843872547149658,
+      "learning_rate": 9.451192254041758e-05,
+      "loss": 0.5859,
+      "step": 160
+    },
+    {
+      "epoch": 0.8947368421052632,
+      "grad_norm": 7.292293548583984,
+      "learning_rate": 9.371733080722911e-05,
+      "loss": 0.4765,
+      "step": 170
+    },
+    {
+      "epoch": 0.9473684210526315,
+      "grad_norm": 2.1483707427978516,
+      "learning_rate": 9.287285680653254e-05,
+      "loss": 0.4496,
+      "step": 180
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 1.2343636751174927,
+      "learning_rate": 9.197946409848194e-05,
+      "loss": 0.42,
+      "step": 190
+    },
+    {
+      "epoch": 1.0526315789473684,
+      "grad_norm": 3.902963638305664,
+      "learning_rate": 9.103817206036382e-05,
+      "loss": 0.4647,
+      "step": 200
+    },
+    {
+      "epoch": 1.0526315789473684,
+      "eval_loss": 0.2513052821159363,
+      "eval_runtime": 0.6934,
+      "eval_samples_per_second": 57.69,
+      "eval_steps_per_second": 14.423,
+      "step": 200
+    },
+    {
+      "epoch": 1.1052631578947367,
+      "grad_norm": 3.0811166763305664,
+      "learning_rate": 9.005005472346924e-05,
+      "loss": 0.3552,
+      "step": 210
+    },
+    {
+      "epoch": 1.1578947368421053,
+      "grad_norm": 12.97741413116455,
+      "learning_rate": 8.90162395476046e-05,
+      "loss": 0.5985,
+      "step": 220
+    },
+    {
+      "epoch": 1.2105263157894737,
+      "grad_norm": 4.255702495574951,
+      "learning_rate": 8.793790613463955e-05,
+      "loss": 0.6665,
+      "step": 230
+    },
+    {
+      "epoch": 1.263157894736842,
+      "grad_norm": 3.5546047687530518,
+      "learning_rate": 8.681628488255986e-05,
+      "loss": 0.4249,
+      "step": 240
+    },
+    {
+      "epoch": 1.3157894736842106,
+      "grad_norm": 9.395575523376465,
+      "learning_rate": 8.565265558156101e-05,
+      "loss": 0.4735,
+      "step": 250
+    },
+    {
+      "epoch": 1.368421052631579,
+      "grad_norm": 3.7864913940429688,
+      "learning_rate": 8.444834595378434e-05,
+      "loss": 0.265,
+      "step": 260
+    },
+    {
+      "epoch": 1.4210526315789473,
+      "grad_norm": 6.043397426605225,
+      "learning_rate": 8.320473013836196e-05,
+      "loss": 0.5925,
+      "step": 270
+    },
+    {
+      "epoch": 1.4736842105263157,
+      "grad_norm": 5.700586795806885,
+      "learning_rate": 8.192322712349917e-05,
+      "loss": 0.4528,
+      "step": 280
+    },
+    {
+      "epoch": 1.526315789473684,
+      "grad_norm": 2.9391582012176514,
+      "learning_rate": 8.060529912738315e-05,
+      "loss": 0.3558,
+      "step": 290
+    },
+    {
+      "epoch": 1.5789473684210527,
+      "grad_norm": 1.8771816492080688,
+      "learning_rate": 7.925244992976538e-05,
+      "loss": 0.5231,
+      "step": 300
+    },
+    {
+      "epoch": 1.5789473684210527,
+      "eval_loss": 0.34315773844718933,
+      "eval_runtime": 0.6944,
+      "eval_samples_per_second": 57.607,
+      "eval_steps_per_second": 14.402,
+      "step": 300
+    },
+    {
+      "epoch": 1.631578947368421,
+      "grad_norm": 4.8421478271484375,
+      "learning_rate": 7.786622315612183e-05,
+      "loss": 0.3751,
+      "step": 310
+    },
+    {
+      "epoch": 1.6842105263157894,
+      "grad_norm": 11.370196342468262,
+      "learning_rate": 7.644820051634812e-05,
+      "loss": 0.3844,
+      "step": 320
+    },
+    {
+      "epoch": 1.736842105263158,
+      "grad_norm": 3.433976173400879,
+      "learning_rate": 7.500000000000001e-05,
+      "loss": 0.5589,
+      "step": 330
+    },
+    {
+      "epoch": 1.7894736842105263,
+      "grad_norm": 6.323594570159912,
+      "learning_rate": 7.35232740301378e-05,
+      "loss": 0.2688,
+      "step": 340
+    },
+    {
+      "epoch": 1.8421052631578947,
+      "grad_norm": 0.5850329399108887,
+      "learning_rate": 7.201970757788172e-05,
+      "loss": 0.2528,
+      "step": 350
+    },
+    {
+      "epoch": 1.8947368421052633,
+      "grad_norm": 1.9679738283157349,
+      "learning_rate": 7.049101623982937e-05,
+      "loss": 0.2398,
+      "step": 360
+    },
+    {
+      "epoch": 1.9473684210526314,
+      "grad_norm": 12.53429889678955,
+      "learning_rate": 6.89389442805288e-05,
+      "loss": 0.4979,
+      "step": 370
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 7.08478307723999,
+      "learning_rate": 6.736526264224101e-05,
+      "loss": 0.5478,
+      "step": 380
+    },
+    {
+      "epoch": 2.0526315789473686,
+      "grad_norm": 9.115711212158203,
+      "learning_rate": 6.577176692426279e-05,
+      "loss": 0.4545,
+      "step": 390
+    },
+    {
+      "epoch": 2.1052631578947367,
+      "grad_norm": 6.18236780166626,
+      "learning_rate": 6.416027533411519e-05,
+      "loss": 0.4381,
+      "step": 400
+    },
+    {
+      "epoch": 2.1052631578947367,
+      "eval_loss": 0.21073441207408905,
+      "eval_runtime": 0.6912,
+      "eval_samples_per_second": 57.873,
+      "eval_steps_per_second": 14.468,
+      "step": 400
+    },
+    {
+      "epoch": 2.1578947368421053,
+      "grad_norm": 1.294372797012329,
+      "learning_rate": 6.253262661293604e-05,
+      "loss": 0.2746,
+      "step": 410
+    },
+    {
+      "epoch": 2.2105263157894735,
+      "grad_norm": 9.00351619720459,
+      "learning_rate": 6.0890677937442574e-05,
+      "loss": 0.2276,
+      "step": 420
+    },
+    {
+      "epoch": 2.263157894736842,
+      "grad_norm": 10.986132621765137,
+      "learning_rate": 5.923630280085948e-05,
+      "loss": 0.2474,
+      "step": 430
+    },
+    {
+      "epoch": 2.3157894736842106,
+      "grad_norm": 12.219366073608398,
+      "learning_rate": 5.757138887522884e-05,
+      "loss": 0.2644,
+      "step": 440
+    },
+    {
+      "epoch": 2.3684210526315788,
+      "grad_norm": 2.5177173614501953,
+      "learning_rate": 5.5897835857542317e-05,
+      "loss": 0.3294,
+      "step": 450
+    },
+    {
+      "epoch": 2.4210526315789473,
+      "grad_norm": 6.946788787841797,
+      "learning_rate": 5.4217553302152237e-05,
+      "loss": 0.5553,
+      "step": 460
+    },
+    {
+      "epoch": 2.473684210526316,
+      "grad_norm": 3.0805630683898926,
+      "learning_rate": 5.2532458441935636e-05,
+      "loss": 0.1107,
+      "step": 470
+    },
+    {
+      "epoch": 2.526315789473684,
+      "grad_norm": 3.5907771587371826,
+      "learning_rate": 5.084447400069655e-05,
+      "loss": 0.2507,
+      "step": 480
+    },
+    {
+      "epoch": 2.5789473684210527,
+      "grad_norm": 1.5780644416809082,
+      "learning_rate": 4.915552599930345e-05,
+      "loss": 0.1839,
+      "step": 490
+    },
+    {
+      "epoch": 2.6315789473684212,
+      "grad_norm": 2.739733934402466,
+      "learning_rate": 4.746754155806437e-05,
+      "loss": 0.2159,
+      "step": 500
+    },
+    {
+      "epoch": 2.6315789473684212,
+      "eval_loss": 0.36658158898353577,
+      "eval_runtime": 0.6913,
+      "eval_samples_per_second": 57.864,
+      "eval_steps_per_second": 14.466,
+      "step": 500
+    },
+    {
+      "epoch": 2.6842105263157894,
+      "grad_norm": 1.3136796951293945,
+      "learning_rate": 4.578244669784777e-05,
+      "loss": 0.1742,
+      "step": 510
+    },
+    {
+      "epoch": 2.736842105263158,
+      "grad_norm": 0.0468708761036396,
+      "learning_rate": 4.410216414245771e-05,
+      "loss": 0.3915,
+      "step": 520
+    },
+    {
+      "epoch": 2.7894736842105265,
+      "grad_norm": 12.833720207214355,
+      "learning_rate": 4.2428611124771184e-05,
+      "loss": 0.2055,
+      "step": 530
+    },
+    {
+      "epoch": 2.8421052631578947,
+      "grad_norm": 0.10741530358791351,
+      "learning_rate": 4.076369719914055e-05,
+      "loss": 0.3395,
+      "step": 540
+    },
+    {
+      "epoch": 2.8947368421052633,
+      "grad_norm": 6.424005508422852,
+      "learning_rate": 3.9109322062557424e-05,
+      "loss": 0.2351,
+      "step": 550
+    },
+    {
+      "epoch": 2.9473684210526314,
+      "grad_norm": 7.805177211761475,
+      "learning_rate": 3.746737338706397e-05,
+      "loss": 0.1548,
+      "step": 560
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 1.9612077474594116,
+      "learning_rate": 3.58397246658848e-05,
+      "loss": 0.1734,
+      "step": 570
+    },
+    {
+      "epoch": 3.0526315789473686,
+      "grad_norm": 6.020046234130859,
+      "learning_rate": 3.422823307573722e-05,
+      "loss": 0.1015,
+      "step": 580
+    },
+    {
+      "epoch": 3.1052631578947367,
+      "grad_norm": 5.7347307205200195,
+      "learning_rate": 3.263473735775899e-05,
+      "loss": 0.1705,
+      "step": 590
+    },
+    {
+      "epoch": 3.1578947368421053,
+      "grad_norm": 15.45874309539795,
+      "learning_rate": 3.10610557194712e-05,
+      "loss": 0.1934,
+      "step": 600
+    },
+    {
+      "epoch": 3.1578947368421053,
+      "eval_loss": 0.26252931356430054,
+      "eval_runtime": 0.6908,
+      "eval_samples_per_second": 57.906,
+      "eval_steps_per_second": 14.476,
+      "step": 600
+    },
+    {
+      "epoch": 3.2105263157894735,
+      "grad_norm": 1.443434715270996,
+      "learning_rate": 2.950898376017064e-05,
+      "loss": 0.1024,
+      "step": 610
+    },
+    {
+      "epoch": 3.263157894736842,
+      "grad_norm": 8.99201488494873,
+      "learning_rate": 2.798029242211828e-05,
+      "loss": 0.1786,
+      "step": 620
+    },
+    {
+      "epoch": 3.3157894736842106,
+      "grad_norm": 3.1811983585357666,
+      "learning_rate": 2.6476725969862227e-05,
+      "loss": 0.1907,
+      "step": 630
+    },
+    {
+      "epoch": 3.3684210526315788,
+      "grad_norm": 16.603193283081055,
+      "learning_rate": 2.500000000000001e-05,
+      "loss": 0.125,
+      "step": 640
+    },
+    {
+      "epoch": 3.4210526315789473,
+      "grad_norm": 6.346674919128418,
+      "learning_rate": 2.3551799483651894e-05,
+      "loss": 0.1509,
+      "step": 650
+    },
+    {
+      "epoch": 3.473684210526316,
+      "grad_norm": 3.2951841354370117,
+      "learning_rate": 2.2133776843878186e-05,
+      "loss": 0.1502,
+      "step": 660
+    },
+    {
+      "epoch": 3.526315789473684,
+      "grad_norm": 4.906289100646973,
+      "learning_rate": 2.074755007023461e-05,
+      "loss": 0.1458,
+      "step": 670
+    },
+    {
+      "epoch": 3.5789473684210527,
+      "grad_norm": 0.9275074005126953,
+      "learning_rate": 1.9394700872616855e-05,
+      "loss": 0.2181,
+      "step": 680
+    },
+    {
+      "epoch": 3.6315789473684212,
+      "grad_norm": 0.8629674315452576,
+      "learning_rate": 1.807677287650083e-05,
+      "loss": 0.0979,
+      "step": 690
+    },
+    {
+      "epoch": 3.6842105263157894,
+      "grad_norm": 6.139326095581055,
+      "learning_rate": 1.6795269861638042e-05,
+      "loss": 0.2701,
+      "step": 700
+    },
+    {
+      "epoch": 3.6842105263157894,
+      "eval_loss": 0.2853814661502838,
+      "eval_runtime": 0.691,
+      "eval_samples_per_second": 57.884,
+      "eval_steps_per_second": 14.471,
+      "step": 700
+    },
+    {
+      "epoch": 3.6842105263157894,
+      "step": 700,
+      "total_flos": 1.461153753759744e+16,
+      "train_loss": 0.5859067852156503,
+      "train_runtime": 192.1858,
+      "train_samples_per_second": 19.773,
+      "train_steps_per_second": 4.943
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 950,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 5,
+  "save_steps": 100,
+  "total_flos": 1.461153753759744e+16,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

llama2_13b_peft/object_counting/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84748a17ab24fc05e783a2f81ad62b8bc6f4c5abe17d42454e72ec6964e7435f
+size 5176