diff --git a/lora-microsoft-llava-med-v1.5-mistral-7b-train_28k_custom-train-mlp-and-llm-unquantized-epoch-3-lr-6e-5/README.md b/lora-microsoft-llava-med-v1.5-mistral-7b-train_28k_custom-train-mlp-and-llm-unquantized-epoch-3-lr-6e-5/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..8ebb3aeb247ddc9079538b09a0e21b51130cf665
--- /dev/null
+++ b/lora-microsoft-llava-med-v1.5-mistral-7b-train_28k_custom-train-mlp-and-llm-unquantized-epoch-3-lr-6e-5/README.md
@@ -0,0 +1,202 @@
+---
+base_model: microsoft/llava-med-v1.5-mistral-7b
+library_name: peft
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.10.0
\ No newline at end of file
diff --git a/lora-microsoft-llava-med-v1.5-mistral-7b-train_28k_custom-train-mlp-and-llm-unquantized-epoch-3-lr-6e-5/adapter_config.json b/lora-microsoft-llava-med-v1.5-mistral-7b-train_28k_custom-train-mlp-and-llm-unquantized-epoch-3-lr-6e-5/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..9ba38f38612f23e5dae53e4f98667e6e97a13d1b
--- /dev/null
+++ b/lora-microsoft-llava-med-v1.5-mistral-7b-train_28k_custom-train-mlp-and-llm-unquantized-epoch-3-lr-6e-5/adapter_config.json
@@ -0,0 +1,34 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "/home/FYP/angk0064/ANGK0064/checkpoints/microsoft-llava-med-v1.5-mistral-7b-train_28k_custom-train-mlp-and-llm-unquantized-epoch-2-lr-6e-5",
+  "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": 256,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 128,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "q_proj",
+    "o_proj",
+    "up_proj",
+    "down_proj",
+    "gate_proj",
+    "k_proj",
+    "v_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}
\ No newline at end of file
diff --git a/lora-microsoft-llava-med-v1.5-mistral-7b-train_28k_custom-train-mlp-and-llm-unquantized-epoch-3-lr-6e-5/adapter_model.safetensors b/lora-microsoft-llava-med-v1.5-mistral-7b-train_28k_custom-train-mlp-and-llm-unquantized-epoch-3-lr-6e-5/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..37e2596f561f2af284cf40d3e5bb84596b5b9805
--- /dev/null
+++ b/lora-microsoft-llava-med-v1.5-mistral-7b-train_28k_custom-train-mlp-and-llm-unquantized-epoch-3-lr-6e-5/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1f80723fc868ee9830de2894900ebd0843444503ce7dcafbd1d46645a4751b0e
+size 671149616
diff --git a/lora-microsoft-llava-med-v1.5-mistral-7b-train_28k_custom-train-mlp-and-llm-unquantized-epoch-3-lr-6e-5/config.json b/lora-microsoft-llava-med-v1.5-mistral-7b-train_28k_custom-train-mlp-and-llm-unquantized-epoch-3-lr-6e-5/config.json
new file mode 100644
index 0000000000000000000000000000000000000000..32e18ab6be292441e4b9f5af0558b451d28cf4d5
--- /dev/null
+++ b/lora-microsoft-llava-med-v1.5-mistral-7b-train_28k_custom-train-mlp-and-llm-unquantized-epoch-3-lr-6e-5/config.json
@@ -0,0 +1,50 @@
+{
+  "_name_or_path": "/home/FYP/angk0064/ANGK0064/checkpoints/microsoft-llava-med-v1.5-mistral-7b-train_28k_custom-train-mlp-and-llm-unquantized-epoch-2-lr-6e-5",
+  "architectures": [
+    "LlavaMistralForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "feature_outs": "encoder+decoder",
+  "freeze_mm_mlp_adapter": false,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "image_aspect_ratio": "pad",
+  "image_processor_path": "openai/clip-vit-large-patch14-336",
+  "img_size": 640,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 32768,
+  "mm_hidden_size": 1024,
+  "mm_projector_lr": 2e-05,
+  "mm_projector_type": "mlp2x_gelu",
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -2,
+  "mm_vision_tower": "/home/FYP/angk0064/ANGK0064/checkpoints/microsoft-llava-med-v1.5-mistral-7b-train_28k_custom-train-mlp-and-llm-unquantized-epoch-2-lr-6e-5",
+  "model_type": "llava_mistral",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "proj_vis_to_txt_tokens": false,
+  "prompt_segtok_w_instruct": false,
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 1000000.0,
+  "segtok_posembed": "sincos",
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "tokenizer_model_max_length": 2048,
+  "tokenizer_padding_side": "right",
+  "torch_dtype": "float16",
+  "transformers_version": "4.37.2",
+  "tune_mm_mlp_adapter": false,
+  "tune_vision_tokenizer": "none",
+  "use_cache": true,
+  "use_mm_proj": true,
+  "vision_backbone": "convnextlarge",
+  "vision_tokenizer_lr": null,
+  "vision_tower_path": "openai/clip-vit-large-patch14-336",
+  "vocab_size": 32000
+}
diff --git a/lora-microsoft-llava-med-v1.5-mistral-7b-train_28k_custom-train-mlp-and-llm-unquantized-epoch-3-lr-6e-5/non_lora_trainables.bin b/lora-microsoft-llava-med-v1.5-mistral-7b-train_28k_custom-train-mlp-and-llm-unquantized-epoch-3-lr-6e-5/non_lora_trainables.bin
new file mode 100644
index 0000000000000000000000000000000000000000..b083a615c81e9f0a195e1bd8af9f262ea0fd87ae
--- /dev/null
+++ b/lora-microsoft-llava-med-v1.5-mistral-7b-train_28k_custom-train-mlp-and-llm-unquantized-epoch-3-lr-6e-5/non_lora_trainables.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:886f239a5293fa8e1b4ed1e0e7d2465e32c85b45370e1a1e5b96e60de350673f
+size 41961648
diff --git a/lora-microsoft-llava-med-v1.5-mistral-7b-train_28k_custom-train-mlp-and-llm-unquantized-epoch-3-lr-6e-5/trainer_state.json b/lora-microsoft-llava-med-v1.5-mistral-7b-train_28k_custom-train-mlp-and-llm-unquantized-epoch-3-lr-6e-5/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..0cd24bdbd4bcda48e491b5a09cc8df8552bf7b12
--- /dev/null
+++ b/lora-microsoft-llava-med-v1.5-mistral-7b-train_28k_custom-train-mlp-and-llm-unquantized-epoch-3-lr-6e-5/trainer_state.json
@@ -0,0 +1,10530 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 1750,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.1320754716981131e-06,
+      "loss": 0.984,
+      "step": 1
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 2.2641509433962262e-06,
+      "loss": 0.9944,
+      "step": 2
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 3.39622641509434e-06,
+      "loss": 0.9645,
+      "step": 3
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.5283018867924524e-06,
+      "loss": 1.0069,
+      "step": 4
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 5.660377358490566e-06,
+      "loss": 0.8821,
+      "step": 5
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 6.79245283018868e-06,
+      "loss": 0.949,
+      "step": 6
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 7.924528301886793e-06,
+      "loss": 0.9754,
+      "step": 7
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 9.056603773584905e-06,
+      "loss": 0.7763,
+      "step": 8
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.0188679245283019e-05,
+      "loss": 0.8808,
+      "step": 9
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.1320754716981132e-05,
+      "loss": 0.9795,
+      "step": 10
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.2452830188679246e-05,
+      "loss": 0.9119,
+      "step": 11
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.358490566037736e-05,
+      "loss": 0.9454,
+      "step": 12
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.4716981132075472e-05,
+      "loss": 0.9241,
+      "step": 13
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.5849056603773586e-05,
+      "loss": 0.9764,
+      "step": 14
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.69811320754717e-05,
+      "loss": 0.9526,
+      "step": 15
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.811320754716981e-05,
+      "loss": 0.896,
+      "step": 16
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.9245283018867924e-05,
+      "loss": 0.8531,
+      "step": 17
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.0377358490566037e-05,
+      "loss": 0.9133,
+      "step": 18
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.150943396226415e-05,
+      "loss": 0.908,
+      "step": 19
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.2641509433962265e-05,
+      "loss": 0.9915,
+      "step": 20
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.377358490566038e-05,
+      "loss": 0.8791,
+      "step": 21
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.4905660377358492e-05,
+      "loss": 0.8353,
+      "step": 22
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.6037735849056606e-05,
+      "loss": 0.8526,
+      "step": 23
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.716981132075472e-05,
+      "loss": 0.7265,
+      "step": 24
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.830188679245283e-05,
+      "loss": 0.7584,
+      "step": 25
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.9433962264150944e-05,
+      "loss": 0.8326,
+      "step": 26
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3.056603773584906e-05,
+      "loss": 0.7301,
+      "step": 27
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3.169811320754717e-05,
+      "loss": 0.8348,
+      "step": 28
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3.2830188679245285e-05,
+      "loss": 0.8063,
+      "step": 29
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3.39622641509434e-05,
+      "loss": 0.8789,
+      "step": 30
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3.5094339622641506e-05,
+      "loss": 0.7554,
+      "step": 31
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3.622641509433962e-05,
+      "loss": 0.8189,
+      "step": 32
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3.735849056603773e-05,
+      "loss": 0.7798,
+      "step": 33
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3.849056603773585e-05,
+      "loss": 0.8047,
+      "step": 34
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 3.962264150943396e-05,
+      "loss": 0.7661,
+      "step": 35
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.0754716981132075e-05,
+      "loss": 0.8636,
+      "step": 36
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.188679245283019e-05,
+      "loss": 0.8524,
+      "step": 37
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.30188679245283e-05,
+      "loss": 0.7508,
+      "step": 38
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.4150943396226416e-05,
+      "loss": 0.8043,
+      "step": 39
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.528301886792453e-05,
+      "loss": 0.772,
+      "step": 40
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.641509433962264e-05,
+      "loss": 0.8261,
+      "step": 41
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.754716981132076e-05,
+      "loss": 0.8324,
+      "step": 42
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.867924528301887e-05,
+      "loss": 0.8211,
+      "step": 43
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9811320754716985e-05,
+      "loss": 0.7814,
+      "step": 44
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.09433962264151e-05,
+      "loss": 0.8136,
+      "step": 45
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.207547169811321e-05,
+      "loss": 0.7939,
+      "step": 46
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.3207547169811326e-05,
+      "loss": 0.681,
+      "step": 47
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.433962264150944e-05,
+      "loss": 0.5892,
+      "step": 48
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.5471698113207547e-05,
+      "loss": 0.6615,
+      "step": 49
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.660377358490566e-05,
+      "loss": 0.6939,
+      "step": 50
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.7735849056603774e-05,
+      "loss": 0.7392,
+      "step": 51
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.886792452830189e-05,
+      "loss": 0.7356,
+      "step": 52
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.886792452830189e-05,
+      "loss": 0.8311,
+      "step": 53
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.886792452830189e-05,
+      "loss": 0.7868,
+      "step": 54
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.886792452830189e-05,
+      "loss": 0.874,
+      "step": 55
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 6e-05,
+      "loss": 0.771,
+      "step": 56
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 6e-05,
+      "loss": 4.2661,
+      "step": 57
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 6e-05,
+      "loss": 4.675,
+      "step": 58
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 6e-05,
+      "loss": 4.8925,
+      "step": 59
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 6e-05,
+      "loss": 4.3031,
+      "step": 60
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 6e-05,
+      "loss": 4.6949,
+      "step": 61
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 6e-05,
+      "loss": 4.2882,
+      "step": 62
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 6e-05,
+      "loss": 4.8062,
+      "step": 63
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.999994859241577e-05,
+      "loss": 5.0871,
+      "step": 64
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.999979436983926e-05,
+      "loss": 0.8247,
+      "step": 65
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.999953733279902e-05,
+      "loss": 0.783,
+      "step": 66
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.999917748217596e-05,
+      "loss": 0.8285,
+      "step": 67
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.999871481920335e-05,
+      "loss": 0.8829,
+      "step": 68
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.999814934546682e-05,
+      "loss": 0.8878,
+      "step": 69
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.999748106290433e-05,
+      "loss": 0.9087,
+      "step": 70
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.999670997380621e-05,
+      "loss": 0.8256,
+      "step": 71
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.999583608081512e-05,
+      "loss": 0.7756,
+      "step": 72
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.999485938692604e-05,
+      "loss": 0.8242,
+      "step": 73
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.9993779895486265e-05,
+      "loss": 0.7965,
+      "step": 74
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.99925976101954e-05,
+      "loss": 0.8755,
+      "step": 75
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.999131253510533e-05,
+      "loss": 0.782,
+      "step": 76
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.998992467462024e-05,
+      "loss": 0.823,
+      "step": 77
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.998843403349657e-05,
+      "loss": 0.7686,
+      "step": 78
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.9986840616843005e-05,
+      "loss": 0.787,
+      "step": 79
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.998514443012045e-05,
+      "loss": 0.7464,
+      "step": 80
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.998334547914202e-05,
+      "loss": 0.7867,
+      "step": 81
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.9981443770073044e-05,
+      "loss": 0.7834,
+      "step": 82
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.997943930943101e-05,
+      "loss": 0.8186,
+      "step": 83
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.9977332104085536e-05,
+      "loss": 0.7469,
+      "step": 84
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.99751221612584e-05,
+      "loss": 0.7891,
+      "step": 85
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.997280948852343e-05,
+      "loss": 0.7547,
+      "step": 86
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.997039409380657e-05,
+      "loss": 0.7452,
+      "step": 87
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.996787598538579e-05,
+      "loss": 0.7281,
+      "step": 88
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.9965255171891076e-05,
+      "loss": 0.719,
+      "step": 89
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.996253166230441e-05,
+      "loss": 0.8765,
+      "step": 90
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.995970546595975e-05,
+      "loss": 0.7934,
+      "step": 91
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.9956776592542924e-05,
+      "loss": 0.8383,
+      "step": 92
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.9953745052091704e-05,
+      "loss": 0.7797,
+      "step": 93
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.99506108549957e-05,
+      "loss": 0.7042,
+      "step": 94
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.994737401199635e-05,
+      "loss": 0.8937,
+      "step": 95
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.994403453418685e-05,
+      "loss": 0.745,
+      "step": 96
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.99405924330122e-05,
+      "loss": 0.8922,
+      "step": 97
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.9937047720269047e-05,
+      "loss": 0.9339,
+      "step": 98
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.9933400408105744e-05,
+      "loss": 0.8125,
+      "step": 99
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.9929650509022254e-05,
+      "loss": 0.773,
+      "step": 100
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.992579803587012e-05,
+      "loss": 0.8901,
+      "step": 101
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.9921843001852445e-05,
+      "loss": 0.796,
+      "step": 102
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.991778542052381e-05,
+      "loss": 0.6898,
+      "step": 103
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.991362530579025e-05,
+      "loss": 0.772,
+      "step": 104
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.990936267190917e-05,
+      "loss": 0.9018,
+      "step": 105
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.990499753348938e-05,
+      "loss": 0.8631,
+      "step": 106
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.990052990549094e-05,
+      "loss": 0.7957,
+      "step": 107
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.9895959803225194e-05,
+      "loss": 0.7115,
+      "step": 108
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.989128724235467e-05,
+      "loss": 0.7696,
+      "step": 109
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.988651223889302e-05,
+      "loss": 0.789,
+      "step": 110
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.988163480920504e-05,
+      "loss": 0.7626,
+      "step": 111
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.987665497000649e-05,
+      "loss": 0.779,
+      "step": 112
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.987157273836414e-05,
+      "loss": 0.7776,
+      "step": 113
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.986638813169568e-05,
+      "loss": 0.7062,
+      "step": 114
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.986110116776966e-05,
+      "loss": 0.7489,
+      "step": 115
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.98557118647054e-05,
+      "loss": 0.752,
+      "step": 116
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.9850220240972986e-05,
+      "loss": 0.7192,
+      "step": 117
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.984462631539314e-05,
+      "loss": 0.6469,
+      "step": 118
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.983893010713722e-05,
+      "loss": 0.7975,
+      "step": 119
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.9833131635727114e-05,
+      "loss": 0.6282,
+      "step": 120
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.982723092103518e-05,
+      "loss": 0.926,
+      "step": 121
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.9821227983284177e-05,
+      "loss": 0.7608,
+      "step": 122
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.981512284304722e-05,
+      "loss": 0.7794,
+      "step": 123
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.980891552124767e-05,
+      "loss": 0.7452,
+      "step": 124
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.9802606039159084e-05,
+      "loss": 0.7814,
+      "step": 125
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.979619441840515e-05,
+      "loss": 0.7865,
+      "step": 126
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.97896806809596e-05,
+      "loss": 0.7994,
+      "step": 127
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.978306484914613e-05,
+      "loss": 0.7883,
+      "step": 128
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.977634694563833e-05,
+      "loss": 0.7384,
+      "step": 129
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.976952699345961e-05,
+      "loss": 0.782,
+      "step": 130
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.976260501598314e-05,
+      "loss": 0.731,
+      "step": 131
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.975558103693171e-05,
+      "loss": 0.8177,
+      "step": 132
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.9748455080377724e-05,
+      "loss": 0.8273,
+      "step": 133
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.974122717074305e-05,
+      "loss": 0.7878,
+      "step": 134
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.973389733279899e-05,
+      "loss": 0.7315,
+      "step": 135
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.972646559166613e-05,
+      "loss": 0.7512,
+      "step": 136
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.9718931972814365e-05,
+      "loss": 0.7192,
+      "step": 137
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.97112965020627e-05,
+      "loss": 0.8082,
+      "step": 138
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.970355920557919e-05,
+      "loss": 0.6569,
+      "step": 139
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.96957201098809e-05,
+      "loss": 0.7265,
+      "step": 140
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.968777924183375e-05,
+      "loss": 0.6206,
+      "step": 141
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.967973662865247e-05,
+      "loss": 0.6747,
+      "step": 142
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.967159229790047e-05,
+      "loss": 0.7292,
+      "step": 143
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.96633462774898e-05,
+      "loss": 0.7897,
+      "step": 144
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.965499859568095e-05,
+      "loss": 0.811,
+      "step": 145
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.96465492810829e-05,
+      "loss": 0.7788,
+      "step": 146
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.96379983626529e-05,
+      "loss": 0.7398,
+      "step": 147
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.9629345869696416e-05,
+      "loss": 0.7591,
+      "step": 148
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.962059183186704e-05,
+      "loss": 0.7619,
+      "step": 149
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.961173627916635e-05,
+      "loss": 0.7297,
+      "step": 150
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.9602779241943856e-05,
+      "loss": 0.6946,
+      "step": 151
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.959372075089688e-05,
+      "loss": 0.716,
+      "step": 152
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.958456083707043e-05,
+      "loss": 0.7513,
+      "step": 153
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.95752995318571e-05,
+      "loss": 0.7938,
+      "step": 154
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.9565936866996975e-05,
+      "loss": 0.7685,
+      "step": 155
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.955647287457753e-05,
+      "loss": 0.7462,
+      "step": 156
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.95469075870335e-05,
+      "loss": 0.6855,
+      "step": 157
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.953724103714676e-05,
+      "loss": 0.7388,
+      "step": 158
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.952747325804625e-05,
+      "loss": 0.7249,
+      "step": 159
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.951760428320784e-05,
+      "loss": 0.6309,
+      "step": 160
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.95076341464542e-05,
+      "loss": 0.7213,
+      "step": 161
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.949756288195469e-05,
+      "loss": 0.7135,
+      "step": 162
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.948739052422529e-05,
+      "loss": 0.7328,
+      "step": 163
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.947711710812842e-05,
+      "loss": 0.6184,
+      "step": 164
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.946674266887284e-05,
+      "loss": 0.6402,
+      "step": 165
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.945626724201354e-05,
+      "loss": 0.7407,
+      "step": 166
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.944569086345162e-05,
+      "loss": 0.693,
+      "step": 167
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.943501356943415e-05,
+      "loss": 0.755,
+      "step": 168
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.942423539655405e-05,
+      "loss": 0.6781,
+      "step": 169
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.9413356381749986e-05,
+      "loss": 0.672,
+      "step": 170
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.94023765623062e-05,
+      "loss": 0.7465,
+      "step": 171
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.9391295975852437e-05,
+      "loss": 0.6744,
+      "step": 172
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.938011466036377e-05,
+      "loss": 0.7075,
+      "step": 173
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.936883265416049e-05,
+      "loss": 0.7211,
+      "step": 174
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.9357449995907996e-05,
+      "loss": 0.7152,
+      "step": 175
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.93459667246166e-05,
+      "loss": 0.6104,
+      "step": 176
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.9334382879641456e-05,
+      "loss": 0.6559,
+      "step": 177
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.932269850068239e-05,
+      "loss": 0.7635,
+      "step": 178
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.93109136277838e-05,
+      "loss": 0.7246,
+      "step": 179
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.929902830133445e-05,
+      "loss": 0.6343,
+      "step": 180
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.9287042562067426e-05,
+      "loss": 0.7213,
+      "step": 181
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.92749564510599e-05,
+      "loss": 0.6929,
+      "step": 182
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 5.926277000973307e-05,
+      "loss": 0.7319,
+      "step": 183
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.925048327985197e-05,
+      "loss": 0.7095,
+      "step": 184
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.9238096303525346e-05,
+      "loss": 0.5801,
+      "step": 185
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.922560912320548e-05,
+      "loss": 0.8254,
+      "step": 186
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.9213021781688104e-05,
+      "loss": 0.8144,
+      "step": 187
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.9200334322112206e-05,
+      "loss": 0.714,
+      "step": 188
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.918754678795989e-05,
+      "loss": 0.7328,
+      "step": 189
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.9174659223056236e-05,
+      "loss": 0.644,
+      "step": 190
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.9161671671569165e-05,
+      "loss": 0.668,
+      "step": 191
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.914858417800923e-05,
+      "loss": 0.7877,
+      "step": 192
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.9135396787229554e-05,
+      "loss": 0.7266,
+      "step": 193
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.912210954442558e-05,
+      "loss": 0.6286,
+      "step": 194
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.910872249513498e-05,
+      "loss": 0.7273,
+      "step": 195
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.9095235685237476e-05,
+      "loss": 0.5564,
+      "step": 196
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.90816491609547e-05,
+      "loss": 0.6883,
+      "step": 197
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.906796296885e-05,
+      "loss": 0.7348,
+      "step": 198
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.905417715582831e-05,
+      "loss": 0.7364,
+      "step": 199
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.9040291769135996e-05,
+      "loss": 0.75,
+      "step": 200
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 5.9026306856360675e-05,
+      "loss": 0.7007,
+      "step": 201
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.901222246543105e-05,
+      "loss": 0.7138,
+      "step": 202
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.8998038644616736e-05,
+      "loss": 0.6187,
+      "step": 203
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.898375544252816e-05,
+      "loss": 0.7636,
+      "step": 204
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.8969372908116296e-05,
+      "loss": 0.7033,
+      "step": 205
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.895489109067258e-05,
+      "loss": 0.6516,
+      "step": 206
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.894031003982868e-05,
+      "loss": 0.7494,
+      "step": 207
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.892562980555638e-05,
+      "loss": 0.7549,
+      "step": 208
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.891085043816737e-05,
+      "loss": 0.6532,
+      "step": 209
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.889597198831308e-05,
+      "loss": 0.6035,
+      "step": 210
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.888099450698455e-05,
+      "loss": 0.7376,
+      "step": 211
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.8865918045512155e-05,
+      "loss": 0.6737,
+      "step": 212
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.8850742655565544e-05,
+      "loss": 0.6679,
+      "step": 213
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.8835468389153386e-05,
+      "loss": 0.6503,
+      "step": 214
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.8820095298623216e-05,
+      "loss": 0.4894,
+      "step": 215
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.8804623436661286e-05,
+      "loss": 0.6626,
+      "step": 216
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.878905285629231e-05,
+      "loss": 0.5449,
+      "step": 217
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.8773383610879364e-05,
+      "loss": 0.6856,
+      "step": 218
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.875761575412365e-05,
+      "loss": 0.5696,
+      "step": 219
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.874174934006432e-05,
+      "loss": 0.6401,
+      "step": 220
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.872578442307832e-05,
+      "loss": 0.5707,
+      "step": 221
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.8709721057880155e-05,
+      "loss": 0.6698,
+      "step": 222
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.869355929952177e-05,
+      "loss": 0.7763,
+      "step": 223
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.867729920339227e-05,
+      "loss": 0.7344,
+      "step": 224
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.866094082521781e-05,
+      "loss": 0.6473,
+      "step": 225
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.8644484221061396e-05,
+      "loss": 0.6822,
+      "step": 226
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.862792944732261e-05,
+      "loss": 0.6705,
+      "step": 227
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.861127656073754e-05,
+      "loss": 0.6624,
+      "step": 228
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.859452561837849e-05,
+      "loss": 0.6099,
+      "step": 229
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.8577676677653816e-05,
+      "loss": 0.7413,
+      "step": 230
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.8560729796307745e-05,
+      "loss": 0.6979,
+      "step": 231
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.8543685032420165e-05,
+      "loss": 0.7245,
+      "step": 232
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.852654244440642e-05,
+      "loss": 0.6479,
+      "step": 233
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.8509302091017096e-05,
+      "loss": 0.7111,
+      "step": 234
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.849196403133787e-05,
+      "loss": 0.7985,
+      "step": 235
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 5.847452832478926e-05,
+      "loss": 0.6946,
+      "step": 236
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.845699503112643e-05,
+      "loss": 0.6524,
+      "step": 237
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.8439364210438996e-05,
+      "loss": 0.6167,
+      "step": 238
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.842163592315082e-05,
+      "loss": 0.6808,
+      "step": 239
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.84038102300198e-05,
+      "loss": 0.6548,
+      "step": 240
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.8385887192137664e-05,
+      "loss": 0.6378,
+      "step": 241
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.8367866870929726e-05,
+      "loss": 0.6342,
+      "step": 242
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.8349749328154755e-05,
+      "loss": 0.5874,
+      "step": 243
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.8331534625904685e-05,
+      "loss": 0.5683,
+      "step": 244
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.831322282660444e-05,
+      "loss": 0.6418,
+      "step": 245
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.82948139930117e-05,
+      "loss": 0.6256,
+      "step": 246
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.827630818821672e-05,
+      "loss": 0.5335,
+      "step": 247
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.825770547564209e-05,
+      "loss": 0.6189,
+      "step": 248
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.82390059190425e-05,
+      "loss": 0.5732,
+      "step": 249
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.822020958250454e-05,
+      "loss": 0.68,
+      "step": 250
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.8201316530446514e-05,
+      "loss": 0.6549,
+      "step": 251
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.818232682761815e-05,
+      "loss": 0.5815,
+      "step": 252
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 5.816324053910045e-05,
+      "loss": 0.5863,
+      "step": 253
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.814405773030539e-05,
+      "loss": 0.6609,
+      "step": 254
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.8124778466975785e-05,
+      "loss": 0.7245,
+      "step": 255
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.810540281518497e-05,
+      "loss": 0.7186,
+      "step": 256
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.808593084133665e-05,
+      "loss": 0.6909,
+      "step": 257
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.806636261216464e-05,
+      "loss": 0.6229,
+      "step": 258
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.804669819473263e-05,
+      "loss": 0.6927,
+      "step": 259
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.8026937656433966e-05,
+      "loss": 0.6648,
+      "step": 260
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.8007081064991405e-05,
+      "loss": 0.6513,
+      "step": 261
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.7987128488456934e-05,
+      "loss": 0.6407,
+      "step": 262
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.796707999521143e-05,
+      "loss": 0.6492,
+      "step": 263
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.7946935653964575e-05,
+      "loss": 0.592,
+      "step": 264
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.792669553375447e-05,
+      "loss": 0.5872,
+      "step": 265
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.790635970394751e-05,
+      "loss": 0.701,
+      "step": 266
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.788592823423809e-05,
+      "loss": 0.733,
+      "step": 267
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.786540119464835e-05,
+      "loss": 0.6803,
+      "step": 268
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.784477865552801e-05,
+      "loss": 0.6519,
+      "step": 269
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.782406068755406e-05,
+      "loss": 0.7718,
+      "step": 270
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 5.780324736173056e-05,
+      "loss": 0.7805,
+      "step": 271
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.7782338749388346e-05,
+      "loss": 0.6972,
+      "step": 272
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.776133492218484e-05,
+      "loss": 0.5952,
+      "step": 273
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.7740235952103775e-05,
+      "loss": 0.7112,
+      "step": 274
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.771904191145495e-05,
+      "loss": 0.6988,
+      "step": 275
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.7697752872874015e-05,
+      "loss": 0.6954,
+      "step": 276
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.767636890932215e-05,
+      "loss": 0.6717,
+      "step": 277
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.7654890094085894e-05,
+      "loss": 0.6917,
+      "step": 278
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.763331650077685e-05,
+      "loss": 0.6994,
+      "step": 279
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.761164820333143e-05,
+      "loss": 0.6163,
+      "step": 280
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.758988527601063e-05,
+      "loss": 0.6589,
+      "step": 281
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.7568027793399745e-05,
+      "loss": 0.6546,
+      "step": 282
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.754607583040813e-05,
+      "loss": 0.7278,
+      "step": 283
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.7524029462268955e-05,
+      "loss": 0.6144,
+      "step": 284
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.750188876453892e-05,
+      "loss": 0.5961,
+      "step": 285
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.7479653813098e-05,
+      "loss": 0.6783,
+      "step": 286
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.745732468414922e-05,
+      "loss": 0.6641,
+      "step": 287
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 5.7434901454218346e-05,
+      "loss": 0.7654,
+      "step": 288
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.741238420015363e-05,
+      "loss": 0.6898,
+      "step": 289
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.7389772999125615e-05,
+      "loss": 0.5695,
+      "step": 290
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.736706792862677e-05,
+      "loss": 0.7286,
+      "step": 291
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.7344269066471264e-05,
+      "loss": 0.6705,
+      "step": 292
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.732137649079475e-05,
+      "loss": 0.6626,
+      "step": 293
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.7298390280054015e-05,
+      "loss": 0.6302,
+      "step": 294
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.727531051302677e-05,
+      "loss": 0.6407,
+      "step": 295
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.7252137268811355e-05,
+      "loss": 0.6594,
+      "step": 296
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.7228870626826466e-05,
+      "loss": 0.6298,
+      "step": 297
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.7205510666810886e-05,
+      "loss": 0.6461,
+      "step": 298
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.718205746882324e-05,
+      "loss": 0.5778,
+      "step": 299
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.715851111324166e-05,
+      "loss": 0.7531,
+      "step": 300
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.713487168076358e-05,
+      "loss": 0.6711,
+      "step": 301
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.7111139252405386e-05,
+      "loss": 0.6579,
+      "step": 302
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.708731390950223e-05,
+      "loss": 0.6604,
+      "step": 303
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.706339573370763e-05,
+      "loss": 0.6629,
+      "step": 304
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.7039384806993325e-05,
+      "loss": 0.6145,
+      "step": 305
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 5.701528121164888e-05,
+      "loss": 0.6972,
+      "step": 306
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.699108503028148e-05,
+      "loss": 0.6954,
+      "step": 307
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.69667963458156e-05,
+      "loss": 0.6443,
+      "step": 308
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.694241524149275e-05,
+      "loss": 0.7057,
+      "step": 309
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.6917941800871174e-05,
+      "loss": 0.6329,
+      "step": 310
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.689337610782556e-05,
+      "loss": 0.6321,
+      "step": 311
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.6868718246546784e-05,
+      "loss": 0.585,
+      "step": 312
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.684396830154159e-05,
+      "loss": 0.7032,
+      "step": 313
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.6819126357632276e-05,
+      "loss": 0.7091,
+      "step": 314
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.679419249995649e-05,
+      "loss": 0.5934,
+      "step": 315
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.676916681396685e-05,
+      "loss": 0.6142,
+      "step": 316
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.674404938543069e-05,
+      "loss": 0.6012,
+      "step": 317
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.671884030042977e-05,
+      "loss": 0.6012,
+      "step": 318
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.6693539645359954e-05,
+      "loss": 0.6252,
+      "step": 319
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.666814750693097e-05,
+      "loss": 0.631,
+      "step": 320
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.6642663972166024e-05,
+      "loss": 0.7406,
+      "step": 321
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.66170891284016e-05,
+      "loss": 0.6488,
+      "step": 322
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.659142306328709e-05,
+      "loss": 0.644,
+      "step": 323
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.656566586478451e-05,
+      "loss": 0.676,
+      "step": 324
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.653981762116823e-05,
+      "loss": 0.6449,
+      "step": 325
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.6513878421024626e-05,
+      "loss": 0.6455,
+      "step": 326
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.648784835325181e-05,
+      "loss": 0.7671,
+      "step": 327
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.64617275070593e-05,
+      "loss": 0.6483,
+      "step": 328
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.643551597196775e-05,
+      "loss": 0.6397,
+      "step": 329
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.64092138378086e-05,
+      "loss": 0.6344,
+      "step": 330
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.63828211947238e-05,
+      "loss": 0.6106,
+      "step": 331
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.635633813316547e-05,
+      "loss": 0.6624,
+      "step": 332
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.632976474389563e-05,
+      "loss": 0.7013,
+      "step": 333
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.630310111798588e-05,
+      "loss": 0.6031,
+      "step": 334
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.627634734681704e-05,
+      "loss": 0.5972,
+      "step": 335
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.624950352207889e-05,
+      "loss": 0.6309,
+      "step": 336
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.622256973576987e-05,
+      "loss": 0.6263,
+      "step": 337
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.619554608019668e-05,
+      "loss": 0.6415,
+      "step": 338
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.6168432647974055e-05,
+      "loss": 0.6006,
+      "step": 339
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.614122953202439e-05,
+      "loss": 0.7122,
+      "step": 340
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 5.611393682557746e-05,
+      "loss": 0.7262,
+      "step": 341
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.6086554622170054e-05,
+      "loss": 0.6598,
+      "step": 342
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.605908301564572e-05,
+      "loss": 0.5706,
+      "step": 343
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.603152210015438e-05,
+      "loss": 0.626,
+      "step": 344
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.600387197015205e-05,
+      "loss": 0.5883,
+      "step": 345
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.597613272040046e-05,
+      "loss": 0.5746,
+      "step": 346
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.594830444596683e-05,
+      "loss": 0.5624,
+      "step": 347
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.592038724222343e-05,
+      "loss": 0.5439,
+      "step": 348
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.5892381204847344e-05,
+      "loss": 0.5579,
+      "step": 349
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.586428642982007e-05,
+      "loss": 0.5372,
+      "step": 350
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.5836103013427235e-05,
+      "loss": 0.5842,
+      "step": 351
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.58078310522583e-05,
+      "loss": 0.6413,
+      "step": 352
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.5779470643206105e-05,
+      "loss": 0.6746,
+      "step": 353
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.5751021883466687e-05,
+      "loss": 0.7076,
+      "step": 354
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.572248487053883e-05,
+      "loss": 0.7278,
+      "step": 355
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.56938597022238e-05,
+      "loss": 0.7096,
+      "step": 356
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.566514647662498e-05,
+      "loss": 0.6924,
+      "step": 357
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 5.563634529214755e-05,
+      "loss": 0.7124,
+      "step": 358
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.560745624749811e-05,
+      "loss": 0.6667,
+      "step": 359
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.5578479441684415e-05,
+      "loss": 0.614,
+      "step": 360
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.5549414974014955e-05,
+      "loss": 0.651,
+      "step": 361
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.552026294409867e-05,
+      "loss": 0.7553,
+      "step": 362
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.549102345184461e-05,
+      "loss": 0.5909,
+      "step": 363
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.546169659746151e-05,
+      "loss": 0.5747,
+      "step": 364
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.54322824814576e-05,
+      "loss": 0.6266,
+      "step": 365
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.540278120464008e-05,
+      "loss": 0.5417,
+      "step": 366
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.537319286811495e-05,
+      "loss": 0.5682,
+      "step": 367
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.534351757328651e-05,
+      "loss": 0.6691,
+      "step": 368
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.5313755421857105e-05,
+      "loss": 0.7497,
+      "step": 369
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.528390651582677e-05,
+      "loss": 0.6403,
+      "step": 370
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.525397095749284e-05,
+      "loss": 0.6983,
+      "step": 371
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.522394884944964e-05,
+      "loss": 0.614,
+      "step": 372
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.519384029458809e-05,
+      "loss": 0.6517,
+      "step": 373
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.5163645396095406e-05,
+      "loss": 0.6558,
+      "step": 374
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.513336425745471e-05,
+      "loss": 0.6424,
+      "step": 375
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 5.5102996982444676e-05,
+      "loss": 0.7108,
+      "step": 376
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.507254367513919e-05,
+      "loss": 0.6603,
+      "step": 377
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.504200443990697e-05,
+      "loss": 0.6206,
+      "step": 378
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.5011379381411254e-05,
+      "loss": 0.6693,
+      "step": 379
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.4980668604609386e-05,
+      "loss": 0.6484,
+      "step": 380
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.4949872214752495e-05,
+      "loss": 0.5783,
+      "step": 381
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.4918990317385095e-05,
+      "loss": 0.684,
+      "step": 382
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.488802301834479e-05,
+      "loss": 0.6076,
+      "step": 383
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.485697042376185e-05,
+      "loss": 0.6367,
+      "step": 384
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.4825832640058855e-05,
+      "loss": 0.7193,
+      "step": 385
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.479460977395035e-05,
+      "loss": 0.6918,
+      "step": 386
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.476330193244249e-05,
+      "loss": 0.6837,
+      "step": 387
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.473190922283264e-05,
+      "loss": 0.6559,
+      "step": 388
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.470043175270901e-05,
+      "loss": 0.7073,
+      "step": 389
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.4668869629950325e-05,
+      "loss": 0.5664,
+      "step": 390
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.4637222962725416e-05,
+      "loss": 0.6026,
+      "step": 391
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.460549185949287e-05,
+      "loss": 0.7343,
+      "step": 392
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 5.457367642900063e-05,
+      "loss": 0.6989,
+      "step": 393
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.454177678028566e-05,
+      "loss": 0.5763,
+      "step": 394
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.4509793022673566e-05,
+      "loss": 0.5589,
+      "step": 395
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.447772526577819e-05,
+      "loss": 0.6426,
+      "step": 396
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.444557361950124e-05,
+      "loss": 0.5671,
+      "step": 397
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.441333819403198e-05,
+      "loss": 0.6228,
+      "step": 398
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.438101909984673e-05,
+      "loss": 0.5886,
+      "step": 399
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.434861644770863e-05,
+      "loss": 0.5241,
+      "step": 400
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.431613034866713e-05,
+      "loss": 0.6338,
+      "step": 401
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.4283560914057696e-05,
+      "loss": 0.5538,
+      "step": 402
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.425090825550138e-05,
+      "loss": 0.6083,
+      "step": 403
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.421817248490448e-05,
+      "loss": 0.5177,
+      "step": 404
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.418535371445812e-05,
+      "loss": 0.5419,
+      "step": 405
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.415245205663788e-05,
+      "loss": 0.6183,
+      "step": 406
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.4119467624203406e-05,
+      "loss": 0.6466,
+      "step": 407
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.408640053019803e-05,
+      "loss": 0.6683,
+      "step": 408
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.405325088794839e-05,
+      "loss": 0.6281,
+      "step": 409
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.402001881106401e-05,
+      "loss": 0.6565,
+      "step": 410
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.398670441343694e-05,
+      "loss": 0.826,
+      "step": 411
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.395330780924137e-05,
+      "loss": 0.6361,
+      "step": 412
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.391982911293322e-05,
+      "loss": 0.6692,
+      "step": 413
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.388626843924974e-05,
+      "loss": 0.6396,
+      "step": 414
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.385262590320914e-05,
+      "loss": 0.6021,
+      "step": 415
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.38189016201102e-05,
+      "loss": 0.6446,
+      "step": 416
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.378509570553183e-05,
+      "loss": 0.6185,
+      "step": 417
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.375120827533274e-05,
+      "loss": 0.554,
+      "step": 418
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.3717239445650984e-05,
+      "loss": 0.6648,
+      "step": 419
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.368318933290358e-05,
+      "loss": 0.6407,
+      "step": 420
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.364905805378615e-05,
+      "loss": 0.5317,
+      "step": 421
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.361484572527247e-05,
+      "loss": 0.6342,
+      "step": 422
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.3580552464614074e-05,
+      "loss": 0.5817,
+      "step": 423
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.3546178389339865e-05,
+      "loss": 0.614,
+      "step": 424
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.3511723617255746e-05,
+      "loss": 0.6035,
+      "step": 425
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.347718826644413e-05,
+      "loss": 0.4957,
+      "step": 426
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.344257245526363e-05,
+      "loss": 0.6218,
+      "step": 427
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 5.340787630234861e-05,
+      "loss": 0.5338,
+      "step": 428
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.337309992660873e-05,
+      "loss": 0.5457,
+      "step": 429
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.333824344722864e-05,
+      "loss": 0.6001,
+      "step": 430
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.330330698366749e-05,
+      "loss": 0.5815,
+      "step": 431
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.326829065565857e-05,
+      "loss": 0.6294,
+      "step": 432
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.323319458320886e-05,
+      "loss": 0.5347,
+      "step": 433
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.319801888659866e-05,
+      "loss": 0.5605,
+      "step": 434
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.3162763686381125e-05,
+      "loss": 0.5103,
+      "step": 435
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.312742910338192e-05,
+      "loss": 0.4829,
+      "step": 436
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.3092015258698736e-05,
+      "loss": 0.6467,
+      "step": 437
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.3056522273700905e-05,
+      "loss": 0.7278,
+      "step": 438
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.302095027002904e-05,
+      "loss": 0.6381,
+      "step": 439
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.298529936959449e-05,
+      "loss": 0.584,
+      "step": 440
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.2949569694579045e-05,
+      "loss": 0.5456,
+      "step": 441
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.2913761367434464e-05,
+      "loss": 0.5252,
+      "step": 442
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.287787451088204e-05,
+      "loss": 0.5806,
+      "step": 443
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.284190924791222e-05,
+      "loss": 0.5732,
+      "step": 444
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.2805865701784164e-05,
+      "loss": 0.6259,
+      "step": 445
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 5.276974399602531e-05,
+      "loss": 0.5319,
+      "step": 446
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.273354425443096e-05,
+      "loss": 0.7098,
+      "step": 447
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.2697266601063864e-05,
+      "loss": 0.6121,
+      "step": 448
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.26609111602538e-05,
+      "loss": 0.6479,
+      "step": 449
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.2624478056597126e-05,
+      "loss": 0.5283,
+      "step": 450
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.258796741495635e-05,
+      "loss": 0.5856,
+      "step": 451
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.255137936045974e-05,
+      "loss": 0.6876,
+      "step": 452
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.251471401850088e-05,
+      "loss": 0.5363,
+      "step": 453
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.247797151473819e-05,
+      "loss": 0.5265,
+      "step": 454
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.2441151975094566e-05,
+      "loss": 0.603,
+      "step": 455
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.2404255525756914e-05,
+      "loss": 0.5615,
+      "step": 456
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.236728229317573e-05,
+      "loss": 0.5103,
+      "step": 457
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.233023240406464e-05,
+      "loss": 0.5616,
+      "step": 458
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.229310598540001e-05,
+      "loss": 0.535,
+      "step": 459
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.225590316442047e-05,
+      "loss": 0.6311,
+      "step": 460
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.221862406862649e-05,
+      "loss": 0.6247,
+      "step": 461
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.218126882577995e-05,
+      "loss": 0.5567,
+      "step": 462
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 5.2143837563903716e-05,
+      "loss": 0.6215,
+      "step": 463
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.210633041128116e-05,
+      "loss": 0.6215,
+      "step": 464
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.206874749645577e-05,
+      "loss": 0.6471,
+      "step": 465
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.2031088948230655e-05,
+      "loss": 0.6479,
+      "step": 466
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.199335489566815e-05,
+      "loss": 0.6193,
+      "step": 467
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.1955545468089364e-05,
+      "loss": 0.4993,
+      "step": 468
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.1917660795073707e-05,
+      "loss": 0.5603,
+      "step": 469
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.187970100645849e-05,
+      "loss": 0.7451,
+      "step": 470
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.184166623233844e-05,
+      "loss": 0.4937,
+      "step": 471
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.18035566030653e-05,
+      "loss": 0.6157,
+      "step": 472
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.17653722492473e-05,
+      "loss": 0.5498,
+      "step": 473
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.172711330174884e-05,
+      "loss": 0.5159,
+      "step": 474
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.168877989168989e-05,
+      "loss": 0.6225,
+      "step": 475
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.165037215044566e-05,
+      "loss": 0.6218,
+      "step": 476
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.16118902096461e-05,
+      "loss": 0.5637,
+      "step": 477
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.157333420117547e-05,
+      "loss": 0.6036,
+      "step": 478
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.153470425717182e-05,
+      "loss": 0.6378,
+      "step": 479
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.149600051002664e-05,
+      "loss": 0.7612,
+      "step": 480
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 5.145722309238434e-05,
+      "loss": 0.6266,
+      "step": 481
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.141837213714181e-05,
+      "loss": 0.6315,
+      "step": 482
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.137944777744796e-05,
+      "loss": 0.6208,
+      "step": 483
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.1340450146703285e-05,
+      "loss": 0.6062,
+      "step": 484
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.130137937855937e-05,
+      "loss": 0.6455,
+      "step": 485
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.1262235606918495e-05,
+      "loss": 0.6741,
+      "step": 486
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.1223018965933086e-05,
+      "loss": 0.664,
+      "step": 487
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.1183729590005335e-05,
+      "loss": 0.6298,
+      "step": 488
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.114436761378671e-05,
+      "loss": 0.5044,
+      "step": 489
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.1104933172177474e-05,
+      "loss": 0.5887,
+      "step": 490
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.1065426400326255e-05,
+      "loss": 0.606,
+      "step": 491
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.1025847433629577e-05,
+      "loss": 0.6814,
+      "step": 492
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.0986196407731356e-05,
+      "loss": 0.6358,
+      "step": 493
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.09464734585225e-05,
+      "loss": 0.5638,
+      "step": 494
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.090667872214041e-05,
+      "loss": 0.5959,
+      "step": 495
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.086681233496849e-05,
+      "loss": 0.6156,
+      "step": 496
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.0826874433635714e-05,
+      "loss": 0.5617,
+      "step": 497
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 5.078686515501616e-05,
+      "loss": 0.6044,
+      "step": 498
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.0746784636228534e-05,
+      "loss": 0.5632,
+      "step": 499
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.070663301463564e-05,
+      "loss": 0.5199,
+      "step": 500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.066641042784403e-05,
+      "loss": 0.5574,
+      "step": 501
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.062611701370344e-05,
+      "loss": 0.5314,
+      "step": 502
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.058575291030633e-05,
+      "loss": 0.6425,
+      "step": 503
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.054531825598744e-05,
+      "loss": 0.5785,
+      "step": 504
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.05048131893233e-05,
+      "loss": 0.5986,
+      "step": 505
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.046423784913175e-05,
+      "loss": 0.6217,
+      "step": 506
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.042359237447146e-05,
+      "loss": 0.5484,
+      "step": 507
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.03828769046415e-05,
+      "loss": 0.6129,
+      "step": 508
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.034209157918077e-05,
+      "loss": 0.615,
+      "step": 509
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.030123653786763e-05,
+      "loss": 0.6129,
+      "step": 510
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.0260311920719326e-05,
+      "loss": 0.615,
+      "step": 511
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.02193178679916e-05,
+      "loss": 0.572,
+      "step": 512
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.01782545201781e-05,
+      "loss": 0.6172,
+      "step": 513
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.013712201801002e-05,
+      "loss": 0.6512,
+      "step": 514
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.00959205024555e-05,
+      "loss": 0.6121,
+      "step": 515
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.005465011471927e-05,
+      "loss": 0.5819,
+      "step": 516
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 5.001331099624204e-05,
+      "loss": 0.4843,
+      "step": 517
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.997190328870009e-05,
+      "loss": 0.4802,
+      "step": 518
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.993042713400477e-05,
+      "loss": 0.5593,
+      "step": 519
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9888882674302004e-05,
+      "loss": 0.6294,
+      "step": 520
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.984727005197182e-05,
+      "loss": 0.6327,
+      "step": 521
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9805589409627837e-05,
+      "loss": 0.4969,
+      "step": 522
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.976384089011679e-05,
+      "loss": 0.6186,
+      "step": 523
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.972202463651808e-05,
+      "loss": 0.5778,
+      "step": 524
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9680140792143174e-05,
+      "loss": 0.5423,
+      "step": 525
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.963818950053525e-05,
+      "loss": 0.4906,
+      "step": 526
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9596170905468594e-05,
+      "loss": 0.5596,
+      "step": 527
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.955408515094817e-05,
+      "loss": 0.5368,
+      "step": 528
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9511932381209126e-05,
+      "loss": 0.5309,
+      "step": 529
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.946971274071625e-05,
+      "loss": 0.5674,
+      "step": 530
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9427426374163535e-05,
+      "loss": 0.5494,
+      "step": 531
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.938507342647364e-05,
+      "loss": 0.631,
+      "step": 532
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.93426540427974e-05,
+      "loss": 0.5246,
+      "step": 533
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.930016836851338e-05,
+      "loss": 0.5858,
+      "step": 534
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.925761654922728e-05,
+      "loss": 0.5856,
+      "step": 535
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.9214998730771526e-05,
+      "loss": 0.6112,
+      "step": 536
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.917231505920472e-05,
+      "loss": 0.5978,
+      "step": 537
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.9129565680811157e-05,
+      "loss": 0.6475,
+      "step": 538
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.908675074210034e-05,
+      "loss": 0.67,
+      "step": 539
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.904387038980642e-05,
+      "loss": 0.5288,
+      "step": 540
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.900092477088775e-05,
+      "loss": 0.5716,
+      "step": 541
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.895791403252637e-05,
+      "loss": 0.5795,
+      "step": 542
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.89148383221275e-05,
+      "loss": 0.5824,
+      "step": 543
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.8871697787319015e-05,
+      "loss": 0.5552,
+      "step": 544
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.882849257595096e-05,
+      "loss": 0.5514,
+      "step": 545
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.8785222836095037e-05,
+      "loss": 0.5995,
+      "step": 546
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.874188871604409e-05,
+      "loss": 0.6254,
+      "step": 547
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.869849036431162e-05,
+      "loss": 0.5405,
+      "step": 548
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.8655027929631265e-05,
+      "loss": 0.5808,
+      "step": 549
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.861150156095627e-05,
+      "loss": 0.6099,
+      "step": 550
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.856791140745899e-05,
+      "loss": 0.4649,
+      "step": 551
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.8524257618530415e-05,
+      "loss": 0.5516,
+      "step": 552
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.8480540343779564e-05,
+      "loss": 0.6138,
+      "step": 553
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.8436759733033105e-05,
+      "loss": 0.5808,
+      "step": 554
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.839291593633471e-05,
+      "loss": 0.6153,
+      "step": 555
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.8349009103944636e-05,
+      "loss": 0.6472,
+      "step": 556
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.830503938633916e-05,
+      "loss": 0.5075,
+      "step": 557
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.8261006934210074e-05,
+      "loss": 0.5973,
+      "step": 558
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.8216911898464184e-05,
+      "loss": 0.6352,
+      "step": 559
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.8172754430222774e-05,
+      "loss": 0.5241,
+      "step": 560
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.812853468082108e-05,
+      "loss": 0.6296,
+      "step": 561
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.808425280180783e-05,
+      "loss": 0.5833,
+      "step": 562
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.8039908944944626e-05,
+      "loss": 0.597,
+      "step": 563
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.799550326220552e-05,
+      "loss": 0.5727,
+      "step": 564
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.7951035905776435e-05,
+      "loss": 0.6011,
+      "step": 565
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.790650702805465e-05,
+      "loss": 0.5076,
+      "step": 566
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.786191678164831e-05,
+      "loss": 0.635,
+      "step": 567
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.7817265319375885e-05,
+      "loss": 0.5674,
+      "step": 568
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.7772552794265606e-05,
+      "loss": 0.6108,
+      "step": 569
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.772777935955501e-05,
+      "loss": 0.6486,
+      "step": 570
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.768294516869037e-05,
+      "loss": 0.5843,
+      "step": 571
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.763805037532619e-05,
+      "loss": 0.6682,
+      "step": 572
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.7593095133324656e-05,
+      "loss": 0.6469,
+      "step": 573
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.7548079596755124e-05,
+      "loss": 0.5807,
+      "step": 574
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.750300391989358e-05,
+      "loss": 0.6577,
+      "step": 575
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.7457868257222165e-05,
+      "loss": 0.5877,
+      "step": 576
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.741267276342856e-05,
+      "loss": 0.6347,
+      "step": 577
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.736741759340549e-05,
+      "loss": 0.633,
+      "step": 578
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.7322102902250237e-05,
+      "loss": 0.5845,
+      "step": 579
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.7276728845264055e-05,
+      "loss": 0.714,
+      "step": 580
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.723129557795165e-05,
+      "loss": 0.6143,
+      "step": 581
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.7185803256020655e-05,
+      "loss": 0.601,
+      "step": 582
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.714025203538109e-05,
+      "loss": 0.4893,
+      "step": 583
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.709464207214485e-05,
+      "loss": 0.5607,
+      "step": 584
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.704897352262513e-05,
+      "loss": 0.5229,
+      "step": 585
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.7003246543335915e-05,
+      "loss": 0.4959,
+      "step": 586
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.695746129099144e-05,
+      "loss": 0.4863,
+      "step": 587
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.6911617922505656e-05,
+      "loss": 0.4787,
+      "step": 588
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.686571659499168e-05,
+      "loss": 0.5803,
+      "step": 589
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.6819757465761284e-05,
+      "loss": 0.4975,
+      "step": 590
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.67737406923243e-05,
+      "loss": 0.6358,
+      "step": 591
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.6727666432388144e-05,
+      "loss": 0.4482,
+      "step": 592
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.668153484385727e-05,
+      "loss": 0.6314,
+      "step": 593
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.6635346084832535e-05,
+      "loss": 0.5261,
+      "step": 594
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.65891003136108e-05,
+      "loss": 0.4681,
+      "step": 595
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.6542797688684294e-05,
+      "loss": 0.5335,
+      "step": 596
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.6496438368740074e-05,
+      "loss": 0.6437,
+      "step": 597
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.645002251265954e-05,
+      "loss": 0.6379,
+      "step": 598
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.64035502795178e-05,
+      "loss": 0.6096,
+      "step": 599
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.6357021828583215e-05,
+      "loss": 0.5648,
+      "step": 600
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.63104373193168e-05,
+      "loss": 0.6142,
+      "step": 601
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.626379691137169e-05,
+      "loss": 0.6019,
+      "step": 602
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.621710076459262e-05,
+      "loss": 0.5394,
+      "step": 603
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.6170349039015315e-05,
+      "loss": 0.521,
+      "step": 604
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.612354189486598e-05,
+      "loss": 0.4978,
+      "step": 605
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.607667949256079e-05,
+      "loss": 0.5498,
+      "step": 606
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.6029761992705244e-05,
+      "loss": 0.6044,
+      "step": 607
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.598278955609371e-05,
+      "loss": 0.6507,
+      "step": 608
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.5935762343708824e-05,
+      "loss": 0.5649,
+      "step": 609
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.588868051672094e-05,
+      "loss": 0.5302,
+      "step": 610
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.584154423648759e-05,
+      "loss": 0.6148,
+      "step": 611
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.579435366455293e-05,
+      "loss": 0.588,
+      "step": 612
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.5747108962647174e-05,
+      "loss": 0.5365,
+      "step": 613
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.569981029268607e-05,
+      "loss": 0.5883,
+      "step": 614
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.5652457816770276e-05,
+      "loss": 0.6049,
+      "step": 615
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.5605051697184925e-05,
+      "loss": 0.6156,
+      "step": 616
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.555759209639894e-05,
+      "loss": 0.5564,
+      "step": 617
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.551007917706453e-05,
+      "loss": 0.5499,
+      "step": 618
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.546251310201668e-05,
+      "loss": 0.6176,
+      "step": 619
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.541489403427252e-05,
+      "loss": 0.5354,
+      "step": 620
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.536722213703079e-05,
+      "loss": 0.537,
+      "step": 621
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.53194975736713e-05,
+      "loss": 0.5231,
+      "step": 622
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.527172050775436e-05,
+      "loss": 0.6024,
+      "step": 623
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.5223891103020194e-05,
+      "loss": 0.6454,
+      "step": 624
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.517600952338841e-05,
+      "loss": 0.6019,
+      "step": 625
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.512807593295744e-05,
+      "loss": 0.6866,
+      "step": 626
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.508009049600395e-05,
+      "loss": 0.6716,
+      "step": 627
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.5032053376982296e-05,
+      "loss": 0.5972,
+      "step": 628
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.498396474052397e-05,
+      "loss": 0.5715,
+      "step": 629
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.4935824751437015e-05,
+      "loss": 0.6071,
+      "step": 630
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.488763357470546e-05,
+      "loss": 0.6668,
+      "step": 631
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.4839391375488776e-05,
+      "loss": 0.5571,
+      "step": 632
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.4791098319121276e-05,
+      "loss": 0.5798,
+      "step": 633
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.474275457111161e-05,
+      "loss": 0.6817,
+      "step": 634
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.4694360297142106e-05,
+      "loss": 0.5551,
+      "step": 635
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.4645915663068294e-05,
+      "loss": 0.5821,
+      "step": 636
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.459742083491828e-05,
+      "loss": 0.6326,
+      "step": 637
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.454887597889218e-05,
+      "loss": 0.4914,
+      "step": 638
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.45002812613616e-05,
+      "loss": 0.5996,
+      "step": 639
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.4451636848869e-05,
+      "loss": 0.7016,
+      "step": 640
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.440294290812715e-05,
+      "loss": 0.527,
+      "step": 641
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.43541996060186e-05,
+      "loss": 0.5022,
+      "step": 642
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.430540710959502e-05,
+      "loss": 0.5501,
+      "step": 643
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.425656558607672e-05,
+      "loss": 0.6487,
+      "step": 644
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.4207675202852e-05,
+      "loss": 0.5827,
+      "step": 645
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.4158736127476635e-05,
+      "loss": 0.7371,
+      "step": 646
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.410974852767326e-05,
+      "loss": 0.5514,
+      "step": 647
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.4060712571330834e-05,
+      "loss": 0.6226,
+      "step": 648
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.4011628426504014e-05,
+      "loss": 0.5673,
+      "step": 649
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.396249626141262e-05,
+      "loss": 0.6267,
+      "step": 650
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.391331624444105e-05,
+      "loss": 0.4739,
+      "step": 651
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.3864088544137705e-05,
+      "loss": 0.7283,
+      "step": 652
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.381481332921438e-05,
+      "loss": 0.6084,
+      "step": 653
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.3765490768545735e-05,
+      "loss": 0.5934,
+      "step": 654
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.3716121031168675e-05,
+      "loss": 0.5977,
+      "step": 655
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.36667042862818e-05,
+      "loss": 0.6399,
+      "step": 656
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.361724070324481e-05,
+      "loss": 0.6026,
+      "step": 657
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.3567730451577925e-05,
+      "loss": 0.5518,
+      "step": 658
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.35181737009613e-05,
+      "loss": 0.6528,
+      "step": 659
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.346857062123447e-05,
+      "loss": 0.5771,
+      "step": 660
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.341892138239573e-05,
+      "loss": 0.6587,
+      "step": 661
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.336922615460157e-05,
+      "loss": 0.6205,
+      "step": 662
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.3319485108166104e-05,
+      "loss": 0.6149,
+      "step": 663
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.326969841356046e-05,
+      "loss": 0.5158,
+      "step": 664
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.3219866241412226e-05,
+      "loss": 0.6378,
+      "step": 665
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.316998876250483e-05,
+      "loss": 0.5595,
+      "step": 666
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.3120066147776994e-05,
+      "loss": 0.6552,
+      "step": 667
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.3070098568322123e-05,
+      "loss": 0.5639,
+      "step": 668
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.302008619538771e-05,
+      "loss": 0.587,
+      "step": 669
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.297002920037479e-05,
+      "loss": 0.9619,
+      "step": 670
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.291992775483729e-05,
+      "loss": 0.6408,
+      "step": 671
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.2869782030481495e-05,
+      "loss": 0.5779,
+      "step": 672
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.281959219916546e-05,
+      "loss": 0.6393,
+      "step": 673
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.276935843289837e-05,
+      "loss": 0.5762,
+      "step": 674
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.271908090384e-05,
+      "loss": 0.5217,
+      "step": 675
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.2668759784300096e-05,
+      "loss": 0.6066,
+      "step": 676
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.2618395246737816e-05,
+      "loss": 0.5595,
+      "step": 677
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.2567987463761104e-05,
+      "loss": 1.2102,
+      "step": 678
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.251753660812612e-05,
+      "loss": 0.5953,
+      "step": 679
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.246704285273662e-05,
+      "loss": 0.581,
+      "step": 680
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.2416506370643436e-05,
+      "loss": 0.5255,
+      "step": 681
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.236592733504376e-05,
+      "loss": 0.5953,
+      "step": 682
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.231530591928069e-05,
+      "loss": 0.5939,
+      "step": 683
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.2264642296842524e-05,
+      "loss": 0.6995,
+      "step": 684
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.221393664136223e-05,
+      "loss": 0.5189,
+      "step": 685
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.216318912661683e-05,
+      "loss": 0.6257,
+      "step": 686
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.211239992652679e-05,
+      "loss": 0.5972,
+      "step": 687
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.206156921515545e-05,
+      "loss": 0.5817,
+      "step": 688
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.201069716670843e-05,
+      "loss": 0.6005,
+      "step": 689
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.195978395553297e-05,
+      "loss": 0.6071,
+      "step": 690
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.190882975611746e-05,
+      "loss": 0.5519,
+      "step": 691
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.1857834743090686e-05,
+      "loss": 0.6594,
+      "step": 692
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.180679909122136e-05,
+      "loss": 0.4901,
+      "step": 693
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.1755722975417444e-05,
+      "loss": 0.4887,
+      "step": 694
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.17046065707256e-05,
+      "loss": 0.5551,
+      "step": 695
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.1653450052330536e-05,
+      "loss": 0.6249,
+      "step": 696
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.160225359555447e-05,
+      "loss": 0.601,
+      "step": 697
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.155101737585646e-05,
+      "loss": 0.6718,
+      "step": 698
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.1499741568831885e-05,
+      "loss": 0.4853,
+      "step": 699
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.144842635021174e-05,
+      "loss": 0.5031,
+      "step": 700
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.139707189586215e-05,
+      "loss": 0.5685,
+      "step": 701
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.1345678381783654e-05,
+      "loss": 0.6181,
+      "step": 702
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.129424598411069e-05,
+      "loss": 0.5952,
+      "step": 703
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.1242774879110945e-05,
+      "loss": 0.5666,
+      "step": 704
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.119126524318475e-05,
+      "loss": 0.5512,
+      "step": 705
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.113971725286453e-05,
+      "loss": 0.5895,
+      "step": 706
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.10881310848141e-05,
+      "loss": 0.5476,
+      "step": 707
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.103650691582816e-05,
+      "loss": 0.5751,
+      "step": 708
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.098484492283162e-05,
+      "loss": 0.625,
+      "step": 709
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.0933145282879046e-05,
+      "loss": 0.6129,
+      "step": 710
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.0881408173154005e-05,
+      "loss": 0.606,
+      "step": 711
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.082963377096847e-05,
+      "loss": 0.489,
+      "step": 712
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.077782225376226e-05,
+      "loss": 0.6747,
+      "step": 713
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.0725973799102364e-05,
+      "loss": 0.5139,
+      "step": 714
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.067408858468236e-05,
+      "loss": 0.4955,
+      "step": 715
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.062216678832183e-05,
+      "loss": 0.5723,
+      "step": 716
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.05702085879657e-05,
+      "loss": 0.5516,
+      "step": 717
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.051821416168368e-05,
+      "loss": 0.5549,
+      "step": 718
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.0466183687669643e-05,
+      "loss": 0.6078,
+      "step": 719
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.0414117344240956e-05,
+      "loss": 0.5591,
+      "step": 720
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.0362015309837976e-05,
+      "loss": 0.5339,
+      "step": 721
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.030987776302334e-05,
+      "loss": 0.5402,
+      "step": 722
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.02577048824814e-05,
+      "loss": 0.598,
+      "step": 723
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.020549684701761e-05,
+      "loss": 0.633,
+      "step": 724
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.0153253835557906e-05,
+      "loss": 0.5791,
+      "step": 725
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.010097602714808e-05,
+      "loss": 0.5982,
+      "step": 726
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.0048663600953187e-05,
+      "loss": 0.6032,
+      "step": 727
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 3.9996316736256934e-05,
+      "loss": 0.5914,
+      "step": 728
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 3.994393561246104e-05,
+      "loss": 0.6243,
+      "step": 729
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 3.989152040908465e-05,
+      "loss": 0.5086,
+      "step": 730
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 3.983907130576367e-05,
+      "loss": 0.5626,
+      "step": 731
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 3.9786588482250236e-05,
+      "loss": 0.6552,
+      "step": 732
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 3.973407211841202e-05,
+      "loss": 0.6599,
+      "step": 733
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 3.9681522394231654e-05,
+      "loss": 0.6344,
+      "step": 734
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 3.962893948980609e-05,
+      "loss": 0.5165,
+      "step": 735
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 3.957632358534599e-05,
+      "loss": 0.5383,
+      "step": 736
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 3.952367486117515e-05,
+      "loss": 0.6451,
+      "step": 737
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 3.947099349772979e-05,
+      "loss": 0.5186,
+      "step": 738
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 3.9418279675558036e-05,
+      "loss": 0.5762,
+      "step": 739
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 3.9365533575319236e-05,
+      "loss": 0.6853,
+      "step": 740
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 3.931275537778336e-05,
+      "loss": 0.5444,
+      "step": 741
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 3.925994526383038e-05,
+      "loss": 0.6784,
+      "step": 742
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 3.9207103414449665e-05,
+      "loss": 0.4994,
+      "step": 743
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 3.915423001073932e-05,
+      "loss": 0.5748,
+      "step": 744
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 3.9101325233905634e-05,
+      "loss": 0.5618,
+      "step": 745
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 3.904838926526236e-05,
+      "loss": 0.5521,
+      "step": 746
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 3.899542228623021e-05,
+      "loss": 0.5654,
+      "step": 747
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 3.8942424478336135e-05,
+      "loss": 0.6504,
+      "step": 748
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 3.8889396023212745e-05,
+      "loss": 0.5636,
+      "step": 749
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 3.88363371025977e-05,
+      "loss": 0.6211,
+      "step": 750
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 3.878324789833307e-05,
+      "loss": 0.5566,
+      "step": 751
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 3.8730128592364686e-05,
+      "loss": 0.556,
+      "step": 752
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 3.867697936674159e-05,
+      "loss": 0.4603,
+      "step": 753
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 3.862380040361531e-05,
+      "loss": 0.5532,
+      "step": 754
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 3.857059188523931e-05,
+      "loss": 0.6093,
+      "step": 755
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 3.8517353993968366e-05,
+      "loss": 0.4688,
+      "step": 756
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 3.8464086912257904e-05,
+      "loss": 0.5026,
+      "step": 757
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 3.841079082266338e-05,
+      "loss": 0.584,
+      "step": 758
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 3.835746590783968e-05,
+      "loss": 0.6131,
+      "step": 759
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 3.830411235054047e-05,
+      "loss": 0.5267,
+      "step": 760
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 3.8250730333617585e-05,
+      "loss": 0.5524,
+      "step": 761
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 3.819732004002039e-05,
+      "loss": 0.579,
+      "step": 762
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 3.8143881652795174e-05,
+      "loss": 0.5172,
+      "step": 763
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 3.809041535508449e-05,
+      "loss": 0.5818,
+      "step": 764
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 3.8036921330126543e-05,
+      "loss": 0.6277,
+      "step": 765
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 3.7983399761254595e-05,
+      "loss": 0.5729,
+      "step": 766
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 3.792985083189625e-05,
+      "loss": 0.5529,
+      "step": 767
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 3.7876274725572947e-05,
+      "loss": 0.577,
+      "step": 768
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 3.782267162589921e-05,
+      "loss": 0.6164,
+      "step": 769
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 3.776904171658211e-05,
+      "loss": 0.5387,
+      "step": 770
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 3.771538518142058e-05,
+      "loss": 0.4804,
+      "step": 771
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 3.7661702204304805e-05,
+      "loss": 0.5999,
+      "step": 772
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 3.7607992969215615e-05,
+      "loss": 0.5258,
+      "step": 773
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 3.7554257660223784e-05,
+      "loss": 0.5418,
+      "step": 774
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 3.75004964614895e-05,
+      "loss": 0.5648,
+      "step": 775
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 3.744670955726165e-05,
+      "loss": 0.6104,
+      "step": 776
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 3.739289713187719e-05,
+      "loss": 0.5385,
+      "step": 777
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 3.7339059369760616e-05,
+      "loss": 0.6153,
+      "step": 778
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 3.7285196455423184e-05,
+      "loss": 0.5865,
+      "step": 779
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 3.723130857346239e-05,
+      "loss": 0.5596,
+      "step": 780
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 3.71773959085613e-05,
+      "loss": 0.6521,
+      "step": 781
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 3.712345864548788e-05,
+      "loss": 0.5737,
+      "step": 782
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 3.7069496969094444e-05,
+      "loss": 0.5223,
+      "step": 783
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 3.7015511064316955e-05,
+      "loss": 0.5533,
+      "step": 784
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 3.696150111617439e-05,
+      "loss": 0.6111,
+      "step": 785
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 3.690746730976817e-05,
+      "loss": 0.4831,
+      "step": 786
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 3.6853409830281434e-05,
+      "loss": 0.5584,
+      "step": 787
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 3.67993288629785e-05,
+      "loss": 0.571,
+      "step": 788
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 3.674522459320414e-05,
+      "loss": 0.5509,
+      "step": 789
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 3.6691097206383024e-05,
+      "loss": 0.5509,
+      "step": 790
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 3.6636946888019036e-05,
+      "loss": 0.5525,
+      "step": 791
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 3.6582773823694626e-05,
+      "loss": 0.5886,
+      "step": 792
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 3.652857819907023e-05,
+      "loss": 0.6096,
+      "step": 793
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 3.6474360199883595e-05,
+      "loss": 0.5614,
+      "step": 794
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 3.642012001194912e-05,
+      "loss": 0.5584,
+      "step": 795
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 3.636585782115732e-05,
+      "loss": 0.6041,
+      "step": 796
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 3.631157381347403e-05,
+      "loss": 0.618,
+      "step": 797
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 3.6257268174939905e-05,
+      "loss": 0.5579,
+      "step": 798
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 3.620294109166974e-05,
+      "loss": 0.5636,
+      "step": 799
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 3.614859274985179e-05,
+      "loss": 0.633,
+      "step": 800
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 3.60942233357472e-05,
+      "loss": 0.5533,
+      "step": 801
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 3.603983303568931e-05,
+      "loss": 0.6926,
+      "step": 802
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 3.598542203608306e-05,
+      "loss": 0.5717,
+      "step": 803
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 3.59309905234043e-05,
+      "loss": 0.5834,
+      "step": 804
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 3.587653868419922e-05,
+      "loss": 0.6262,
+      "step": 805
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 3.582206670508364e-05,
+      "loss": 0.4849,
+      "step": 806
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 3.576757477274243e-05,
+      "loss": 0.5338,
+      "step": 807
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 3.5713063073928826e-05,
+      "loss": 0.5988,
+      "step": 808
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 3.56585317954638e-05,
+      "loss": 0.6399,
+      "step": 809
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 3.560398112423545e-05,
+      "loss": 0.7014,
+      "step": 810
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 3.554941124719833e-05,
+      "loss": 0.6839,
+      "step": 811
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 3.5494822351372805e-05,
+      "loss": 0.5146,
+      "step": 812
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 3.544021462384441e-05,
+      "loss": 0.6736,
+      "step": 813
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 3.538558825176326e-05,
+      "loss": 0.6346,
+      "step": 814
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 3.533094342234333e-05,
+      "loss": 0.5888,
+      "step": 815
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 3.527628032286185e-05,
+      "loss": 0.644,
+      "step": 816
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 3.522159914065871e-05,
+      "loss": 0.5838,
+      "step": 817
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 3.516690006313572e-05,
+      "loss": 0.6062,
+      "step": 818
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 3.5112183277756055e-05,
+      "loss": 0.5796,
+      "step": 819
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 3.5057448972043564e-05,
+      "loss": 0.6718,
+      "step": 820
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 3.500269733358213e-05,
+      "loss": 0.6029,
+      "step": 821
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 3.494792855001506e-05,
+      "loss": 0.6085,
+      "step": 822
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 3.489314280904441e-05,
+      "loss": 0.4753,
+      "step": 823
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 3.483834029843036e-05,
+      "loss": 0.5126,
+      "step": 824
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 3.478352120599054e-05,
+      "loss": 0.5713,
+      "step": 825
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 3.472868571959943e-05,
+      "loss": 0.5892,
+      "step": 826
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 3.467383402718769e-05,
+      "loss": 0.6081,
+      "step": 827
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 3.4618966316741535e-05,
+      "loss": 0.443,
+      "step": 828
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 3.456408277630203e-05,
+      "loss": 0.624,
+      "step": 829
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 3.450918359396455e-05,
+      "loss": 0.6404,
+      "step": 830
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 3.445426895787804e-05,
+      "loss": 0.4569,
+      "step": 831
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 3.439933905624443e-05,
+      "loss": 0.5916,
+      "step": 832
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 3.4344394077317946e-05,
+      "loss": 0.5078,
+      "step": 833
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 3.4289434209404485e-05,
+      "loss": 0.5602,
+      "step": 834
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 3.4234459640861015e-05,
+      "loss": 0.5245,
+      "step": 835
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 3.4179470560094815e-05,
+      "loss": 0.5988,
+      "step": 836
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 3.4124467155562966e-05,
+      "loss": 0.7351,
+      "step": 837
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 3.406944961577161e-05,
+      "loss": 0.6022,
+      "step": 838
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 3.4014418129275314e-05,
+      "loss": 0.5472,
+      "step": 839
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 3.395937288467649e-05,
+      "loss": 0.5789,
+      "step": 840
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 3.390431407062465e-05,
+      "loss": 0.6524,
+      "step": 841
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 3.384924187581585e-05,
+      "loss": 0.5922,
+      "step": 842
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 3.3794156488991994e-05,
+      "loss": 0.6776,
+      "step": 843
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 3.373905809894018e-05,
+      "loss": 0.6063,
+      "step": 844
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 3.3683946894492085e-05,
+      "loss": 0.6022,
+      "step": 845
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 3.362882306452331e-05,
+      "loss": 0.6246,
+      "step": 846
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 3.35736867979527e-05,
+      "loss": 0.6094,
+      "step": 847
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 3.3518538283741765e-05,
+      "loss": 0.5904,
+      "step": 848
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 3.346337771089395e-05,
+      "loss": 0.6046,
+      "step": 849
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 3.340820526845404e-05,
+      "loss": 0.6785,
+      "step": 850
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 3.33530211455075e-05,
+      "loss": 0.6299,
+      "step": 851
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 3.329782553117983e-05,
+      "loss": 0.636,
+      "step": 852
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 3.324261861463591e-05,
+      "loss": 0.628,
+      "step": 853
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 3.318740058507934e-05,
+      "loss": 0.708,
+      "step": 854
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 3.3132171631751855e-05,
+      "loss": 0.5655,
+      "step": 855
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 3.307693194393256e-05,
+      "loss": 0.6016,
+      "step": 856
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 3.3021681710937396e-05,
+      "loss": 0.6359,
+      "step": 857
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 3.296642112211842e-05,
+      "loss": 0.5095,
+      "step": 858
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 3.29111503668632e-05,
+      "loss": 0.5293,
+      "step": 859
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 3.285586963459414e-05,
+      "loss": 0.6051,
+      "step": 860
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 3.280057911476782e-05,
+      "loss": 0.5556,
+      "step": 861
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 3.274527899687439e-05,
+      "loss": 0.5741,
+      "step": 862
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 3.2689969470436883e-05,
+      "loss": 0.6159,
+      "step": 863
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 3.2634650725010544e-05,
+      "loss": 0.6166,
+      "step": 864
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 3.257932295018228e-05,
+      "loss": 0.6688,
+      "step": 865
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 3.25239863355699e-05,
+      "loss": 0.5219,
+      "step": 866
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 3.2468641070821504e-05,
+      "loss": 0.6606,
+      "step": 867
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 3.2413287345614855e-05,
+      "loss": 0.6414,
+      "step": 868
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 3.235792534965671e-05,
+      "loss": 0.6761,
+      "step": 869
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 3.230255527268216e-05,
+      "loss": 0.6125,
+      "step": 870
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 3.2247177304454015e-05,
+      "loss": 0.6975,
+      "step": 871
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 3.219179163476208e-05,
+      "loss": 0.641,
+      "step": 872
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 3.213639845342262e-05,
+      "loss": 0.6408,
+      "step": 873
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 3.2080997950277595e-05,
+      "loss": 0.6033,
+      "step": 874
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 3.202559031519407e-05,
+      "loss": 0.6162,
+      "step": 875
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 3.197017573806357e-05,
+      "loss": 0.5913,
+      "step": 876
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 3.191475440880138e-05,
+      "loss": 0.6345,
+      "step": 877
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 3.185932651734596e-05,
+      "loss": 0.5666,
+      "step": 878
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 3.180389225365823e-05,
+      "loss": 0.5774,
+      "step": 879
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 3.174845180772098e-05,
+      "loss": 0.6279,
+      "step": 880
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 3.169300536953815e-05,
+      "loss": 0.6131,
+      "step": 881
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 3.163755312913424e-05,
+      "loss": 0.5771,
+      "step": 882
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 3.1582095276553645e-05,
+      "loss": 0.6654,
+      "step": 883
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.152663200185997e-05,
+      "loss": 0.6157,
+      "step": 884
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.1471163495135396e-05,
+      "loss": 0.6257,
+      "step": 885
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.1415689946480085e-05,
+      "loss": 0.6467,
+      "step": 886
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.136021154601143e-05,
+      "loss": 0.5768,
+      "step": 887
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.130472848386346e-05,
+      "loss": 0.5877,
+      "step": 888
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.124924095018621e-05,
+      "loss": 0.6378,
+      "step": 889
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.119374913514498e-05,
+      "loss": 0.5318,
+      "step": 890
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.113825322891982e-05,
+      "loss": 0.573,
+      "step": 891
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.108275342170474e-05,
+      "loss": 0.6848,
+      "step": 892
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.1027249903707136e-05,
+      "loss": 0.6087,
+      "step": 893
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.097174286514714e-05,
+      "loss": 0.634,
+      "step": 894
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.0916232496256934e-05,
+      "loss": 0.572,
+      "step": 895
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.086071898728011e-05,
+      "loss": 0.5581,
+      "step": 896
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.080520252847104e-05,
+      "loss": 0.5794,
+      "step": 897
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.074968331009417e-05,
+      "loss": 0.6617,
+      "step": 898
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.069416152242345e-05,
+      "loss": 0.6201,
+      "step": 899
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.063863735574159e-05,
+      "loss": 0.7039,
+      "step": 900
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.058311100033949e-05,
+      "loss": 0.5575,
+      "step": 901
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 3.052758264651553e-05,
+      "loss": 0.6496,
+      "step": 902
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 3.0472052484574953e-05,
+      "loss": 0.494,
+      "step": 903
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 3.0416520704829184e-05,
+      "loss": 0.5869,
+      "step": 904
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 3.0360987497595198e-05,
+      "loss": 0.7095,
+      "step": 905
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 3.0305453053194866e-05,
+      "loss": 0.475,
+      "step": 906
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 3.0249917561954298e-05,
+      "loss": 0.5231,
+      "step": 907
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 3.0194381214203185e-05,
+      "loss": 0.6357,
+      "step": 908
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 3.013884420027417e-05,
+      "loss": 0.5889,
+      "step": 909
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 3.008330671050216e-05,
+      "loss": 0.5796,
+      "step": 910
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 3.0027768935223693e-05,
+      "loss": 0.5547,
+      "step": 911
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 2.9972231064776308e-05,
+      "loss": 0.5914,
+      "step": 912
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 2.9916693289497852e-05,
+      "loss": 0.6359,
+      "step": 913
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 2.986115579972584e-05,
+      "loss": 0.5683,
+      "step": 914
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 2.980561878579681e-05,
+      "loss": 0.6451,
+      "step": 915
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 2.9750082438045703e-05,
+      "loss": 0.6216,
+      "step": 916
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 2.9694546946805142e-05,
+      "loss": 0.6191,
+      "step": 917
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 2.963901250240481e-05,
+      "loss": 0.6004,
+      "step": 918
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.9583479295170824e-05,
+      "loss": 0.6015,
+      "step": 919
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.952794751542505e-05,
+      "loss": 0.5787,
+      "step": 920
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.9472417353484473e-05,
+      "loss": 0.6649,
+      "step": 921
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.9416888999660515e-05,
+      "loss": 0.6742,
+      "step": 922
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.9361362644258416e-05,
+      "loss": 0.6411,
+      "step": 923
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.9305838477576566e-05,
+      "loss": 0.7022,
+      "step": 924
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.925031668990583e-05,
+      "loss": 0.6987,
+      "step": 925
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.9194797471528966e-05,
+      "loss": 0.5453,
+      "step": 926
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.9139281012719892e-05,
+      "loss": 0.638,
+      "step": 927
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.908376750374307e-05,
+      "loss": 0.5711,
+      "step": 928
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.9028257134852867e-05,
+      "loss": 0.6825,
+      "step": 929
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.8972750096292862e-05,
+      "loss": 0.6586,
+      "step": 930
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.8917246578295266e-05,
+      "loss": 0.5799,
+      "step": 931
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.8861746771080185e-05,
+      "loss": 0.6412,
+      "step": 932
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.880625086485502e-05,
+      "loss": 0.6476,
+      "step": 933
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.8750759049813805e-05,
+      "loss": 0.5736,
+      "step": 934
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.8695271516136535e-05,
+      "loss": 0.6578,
+      "step": 935
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 2.8639788453988572e-05,
+      "loss": 0.6835,
+      "step": 936
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.8584310053519916e-05,
+      "loss": 0.6415,
+      "step": 937
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.8528836504864605e-05,
+      "loss": 0.5827,
+      "step": 938
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.8473367998140044e-05,
+      "loss": 0.686,
+      "step": 939
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.8417904723446363e-05,
+      "loss": 0.5839,
+      "step": 940
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.836244687086576e-05,
+      "loss": 0.6204,
+      "step": 941
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.8306994630461858e-05,
+      "loss": 0.7794,
+      "step": 942
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.825154819227903e-05,
+      "loss": 0.5452,
+      "step": 943
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.8196107746341775e-05,
+      "loss": 0.6638,
+      "step": 944
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.8140673482654036e-05,
+      "loss": 0.6623,
+      "step": 945
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.808524559119862e-05,
+      "loss": 0.6876,
+      "step": 946
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.8029824261936438e-05,
+      "loss": 0.5089,
+      "step": 947
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.7974409684805937e-05,
+      "loss": 0.7057,
+      "step": 948
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.791900204972242e-05,
+      "loss": 0.6451,
+      "step": 949
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.7863601546577393e-05,
+      "loss": 0.5296,
+      "step": 950
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.7808208365237914e-05,
+      "loss": 0.6168,
+      "step": 951
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.7752822695545994e-05,
+      "loss": 0.5399,
+      "step": 952
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 2.769744472731784e-05,
+      "loss": 0.5613,
+      "step": 953
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.7642074650343298e-05,
+      "loss": 0.7027,
+      "step": 954
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.7586712654385147e-05,
+      "loss": 0.6108,
+      "step": 955
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.7531358929178498e-05,
+      "loss": 0.5346,
+      "step": 956
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.7476013664430108e-05,
+      "loss": 0.5951,
+      "step": 957
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.7420677049817725e-05,
+      "loss": 0.5667,
+      "step": 958
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.7365349274989457e-05,
+      "loss": 0.6802,
+      "step": 959
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.7310030529563135e-05,
+      "loss": 0.5794,
+      "step": 960
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.7254721003125612e-05,
+      "loss": 0.5452,
+      "step": 961
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.7199420885232175e-05,
+      "loss": 0.5238,
+      "step": 962
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.714413036540586e-05,
+      "loss": 0.5794,
+      "step": 963
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.70888496331368e-05,
+      "loss": 0.5826,
+      "step": 964
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.7033578877881587e-05,
+      "loss": 0.5964,
+      "step": 965
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.6978318289062605e-05,
+      "loss": 0.5782,
+      "step": 966
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.692306805606744e-05,
+      "loss": 0.5367,
+      "step": 967
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.686782836824815e-05,
+      "loss": 0.6902,
+      "step": 968
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.6812599414920657e-05,
+      "loss": 0.5592,
+      "step": 969
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.67573813853641e-05,
+      "loss": 0.6661,
+      "step": 970
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 2.6702174468820172e-05,
+      "loss": 0.6661,
+      "step": 971
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.6646978854492506e-05,
+      "loss": 0.7265,
+      "step": 972
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.659179473154597e-05,
+      "loss": 0.6214,
+      "step": 973
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.653662228910606e-05,
+      "loss": 0.6775,
+      "step": 974
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.648146171625824e-05,
+      "loss": 0.5885,
+      "step": 975
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.6426313202047296e-05,
+      "loss": 0.5505,
+      "step": 976
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.63711769354767e-05,
+      "loss": 0.6895,
+      "step": 977
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.631605310550792e-05,
+      "loss": 0.6386,
+      "step": 978
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.6260941901059826e-05,
+      "loss": 0.5552,
+      "step": 979
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.6205843511008014e-05,
+      "loss": 0.6799,
+      "step": 980
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.615075812418415e-05,
+      "loss": 0.6829,
+      "step": 981
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.6095685929375348e-05,
+      "loss": 0.7311,
+      "step": 982
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.6040627115323515e-05,
+      "loss": 0.5847,
+      "step": 983
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.5985581870724688e-05,
+      "loss": 0.5913,
+      "step": 984
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.59305503842284e-05,
+      "loss": 0.5921,
+      "step": 985
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.5875532844437043e-05,
+      "loss": 0.6697,
+      "step": 986
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.5820529439905186e-05,
+      "loss": 0.7255,
+      "step": 987
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 2.5765540359138993e-05,
+      "loss": 0.533,
+      "step": 988
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.5710565790595516e-05,
+      "loss": 0.7015,
+      "step": 989
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.5655605922682065e-05,
+      "loss": 0.6659,
+      "step": 990
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.560066094375558e-05,
+      "loss": 0.6137,
+      "step": 991
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.5545731042121952e-05,
+      "loss": 0.5658,
+      "step": 992
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.5490816406035445e-05,
+      "loss": 0.6588,
+      "step": 993
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.543591722369797e-05,
+      "loss": 0.6385,
+      "step": 994
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.5381033683258473e-05,
+      "loss": 0.6958,
+      "step": 995
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.532616597281231e-05,
+      "loss": 0.5957,
+      "step": 996
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.5271314280400572e-05,
+      "loss": 0.7036,
+      "step": 997
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.5216478794009465e-05,
+      "loss": 0.6197,
+      "step": 998
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.516165970156965e-05,
+      "loss": 0.5207,
+      "step": 999
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.5106857190955596e-05,
+      "loss": 0.5704,
+      "step": 1000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.5052071449984946e-05,
+      "loss": 0.6606,
+      "step": 1001
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.499730266641787e-05,
+      "loss": 0.6695,
+      "step": 1002
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.494255102795644e-05,
+      "loss": 0.6529,
+      "step": 1003
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.4887816722243947e-05,
+      "loss": 0.6699,
+      "step": 1004
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.483309993686428e-05,
+      "loss": 0.6777,
+      "step": 1005
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 2.4778400859341298e-05,
+      "loss": 0.6526,
+      "step": 1006
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.4723719677138145e-05,
+      "loss": 0.7258,
+      "step": 1007
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.4669056577656677e-05,
+      "loss": 0.6576,
+      "step": 1008
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.4614411748236748e-05,
+      "loss": 0.7345,
+      "step": 1009
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.4559785376155593e-05,
+      "loss": 0.6594,
+      "step": 1010
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.450517764862721e-05,
+      "loss": 0.6993,
+      "step": 1011
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.4450588752801667e-05,
+      "loss": 0.551,
+      "step": 1012
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.4396018875764546e-05,
+      "loss": 0.5206,
+      "step": 1013
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.4341468204536203e-05,
+      "loss": 0.6683,
+      "step": 1014
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.4286936926071182e-05,
+      "loss": 0.6188,
+      "step": 1015
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.4232425227257575e-05,
+      "loss": 0.6339,
+      "step": 1016
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.4177933294916362e-05,
+      "loss": 0.5589,
+      "step": 1017
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.4123461315800785e-05,
+      "loss": 0.6835,
+      "step": 1018
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.4069009476595706e-05,
+      "loss": 0.6139,
+      "step": 1019
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.4014577963916954e-05,
+      "loss": 0.6822,
+      "step": 1020
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.3960166964310698e-05,
+      "loss": 0.5206,
+      "step": 1021
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.3905776664252814e-05,
+      "loss": 0.6279,
+      "step": 1022
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 2.385140725014821e-05,
+      "loss": 0.6376,
+      "step": 1023
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.3797058908330263e-05,
+      "loss": 0.5983,
+      "step": 1024
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.37427318250601e-05,
+      "loss": 0.6465,
+      "step": 1025
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.3688426186525982e-05,
+      "loss": 0.6528,
+      "step": 1026
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.3634142178842695e-05,
+      "loss": 0.6024,
+      "step": 1027
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.357987998805087e-05,
+      "loss": 0.6903,
+      "step": 1028
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.3525639800116413e-05,
+      "loss": 0.6347,
+      "step": 1029
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.3471421800929775e-05,
+      "loss": 0.6623,
+      "step": 1030
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.341722617630538e-05,
+      "loss": 0.6265,
+      "step": 1031
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.336305311198097e-05,
+      "loss": 0.6592,
+      "step": 1032
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.3308902793616968e-05,
+      "loss": 0.7552,
+      "step": 1033
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.3254775406795855e-05,
+      "loss": 0.6032,
+      "step": 1034
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.32006711370215e-05,
+      "loss": 0.753,
+      "step": 1035
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.3146590169718568e-05,
+      "loss": 0.6666,
+      "step": 1036
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.3092532690231845e-05,
+      "loss": 0.5765,
+      "step": 1037
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.3038498883825606e-05,
+      "loss": 0.7197,
+      "step": 1038
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.298448893568305e-05,
+      "loss": 0.7756,
+      "step": 1039
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.2930503030905557e-05,
+      "loss": 0.6007,
+      "step": 1040
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 2.2876541354512124e-05,
+      "loss": 0.7076,
+      "step": 1041
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.2822604091438716e-05,
+      "loss": 0.649,
+      "step": 1042
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.276869142653761e-05,
+      "loss": 0.7868,
+      "step": 1043
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.271480354457682e-05,
+      "loss": 0.6639,
+      "step": 1044
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.2660940630239392e-05,
+      "loss": 0.7125,
+      "step": 1045
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.2607102868122812e-05,
+      "loss": 0.5934,
+      "step": 1046
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.2553290442738367e-05,
+      "loss": 0.6349,
+      "step": 1047
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.2499503538510495e-05,
+      "loss": 0.7039,
+      "step": 1048
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.244574233977621e-05,
+      "loss": 0.6666,
+      "step": 1049
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.2392007030784394e-05,
+      "loss": 0.6505,
+      "step": 1050
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.23382977956952e-05,
+      "loss": 0.6909,
+      "step": 1051
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.2284614818579425e-05,
+      "loss": 0.7319,
+      "step": 1052
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.2230958283417895e-05,
+      "loss": 0.7018,
+      "step": 1053
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.2177328374100794e-05,
+      "loss": 0.6529,
+      "step": 1054
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.212372527442706e-05,
+      "loss": 0.6766,
+      "step": 1055
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.2070149168103753e-05,
+      "loss": 0.656,
+      "step": 1056
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.201660023874542e-05,
+      "loss": 0.6706,
+      "step": 1057
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 2.1963078669873465e-05,
+      "loss": 0.6031,
+      "step": 1058
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.190958464491551e-05,
+      "loss": 0.6684,
+      "step": 1059
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.1856118347204827e-05,
+      "loss": 0.5866,
+      "step": 1060
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.1802679959979613e-05,
+      "loss": 0.6689,
+      "step": 1061
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.1749269666382426e-05,
+      "loss": 0.7511,
+      "step": 1062
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.1695887649459546e-05,
+      "loss": 0.6821,
+      "step": 1063
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.1642534092160323e-05,
+      "loss": 0.642,
+      "step": 1064
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.1589209177336625e-05,
+      "loss": 0.5244,
+      "step": 1065
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.1535913087742104e-05,
+      "loss": 0.6567,
+      "step": 1066
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.1482646006031636e-05,
+      "loss": 0.6817,
+      "step": 1067
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.1429408114760696e-05,
+      "loss": 0.7169,
+      "step": 1068
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.1376199596384698e-05,
+      "loss": 0.6737,
+      "step": 1069
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.1323020633258417e-05,
+      "loss": 0.5672,
+      "step": 1070
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.1269871407635312e-05,
+      "loss": 0.7286,
+      "step": 1071
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.1216752101666934e-05,
+      "loss": 0.6185,
+      "step": 1072
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.1163662897402305e-05,
+      "loss": 0.6989,
+      "step": 1073
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.111060397678726e-05,
+      "loss": 0.6635,
+      "step": 1074
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.1057575521663873e-05,
+      "loss": 0.7397,
+      "step": 1075
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.1004577713769795e-05,
+      "loss": 0.5889,
+      "step": 1076
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.0951610734737648e-05,
+      "loss": 0.6393,
+      "step": 1077
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.0898674766094385e-05,
+      "loss": 0.7026,
+      "step": 1078
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.0845769989260673e-05,
+      "loss": 0.66,
+      "step": 1079
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.0792896585550337e-05,
+      "loss": 0.5725,
+      "step": 1080
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.0740054736169625e-05,
+      "loss": 0.6297,
+      "step": 1081
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.068724462221665e-05,
+      "loss": 0.7333,
+      "step": 1082
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.0634466424680772e-05,
+      "loss": 0.6535,
+      "step": 1083
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.058172032444196e-05,
+      "loss": 0.625,
+      "step": 1084
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.052900650227021e-05,
+      "loss": 0.6404,
+      "step": 1085
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.0476325138824862e-05,
+      "loss": 0.6834,
+      "step": 1086
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.0423676414654015e-05,
+      "loss": 0.6251,
+      "step": 1087
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.0371060510193924e-05,
+      "loss": 0.7271,
+      "step": 1088
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.0318477605768347e-05,
+      "loss": 0.7412,
+      "step": 1089
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.0265927881587977e-05,
+      "loss": 0.7429,
+      "step": 1090
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.021341151774977e-05,
+      "loss": 0.644,
+      "step": 1091
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.016092869423633e-05,
+      "loss": 0.6157,
+      "step": 1092
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 2.0108479590915366e-05,
+      "loss": 0.622,
+      "step": 1093
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 2.0056064387538967e-05,
+      "loss": 0.7285,
+      "step": 1094
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 2.0003683263743064e-05,
+      "loss": 0.6428,
+      "step": 1095
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.9951336399046815e-05,
+      "loss": 0.6739,
+      "step": 1096
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.989902397285193e-05,
+      "loss": 0.7277,
+      "step": 1097
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.9846746164442106e-05,
+      "loss": 0.5932,
+      "step": 1098
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.9794503152982397e-05,
+      "loss": 0.7204,
+      "step": 1099
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.97422951175186e-05,
+      "loss": 0.7261,
+      "step": 1100
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.9690122236976664e-05,
+      "loss": 0.7918,
+      "step": 1101
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.9637984690162026e-05,
+      "loss": 0.7008,
+      "step": 1102
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.9585882655759043e-05,
+      "loss": 0.7094,
+      "step": 1103
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.9533816312330368e-05,
+      "loss": 0.5995,
+      "step": 1104
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.9481785838316314e-05,
+      "loss": 0.7737,
+      "step": 1105
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.9429791412034302e-05,
+      "loss": 0.7207,
+      "step": 1106
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.9377833211678178e-05,
+      "loss": 0.7113,
+      "step": 1107
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.932591141531764e-05,
+      "loss": 0.6705,
+      "step": 1108
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.927402620089764e-05,
+      "loss": 0.7633,
+      "step": 1109
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.9222177746237742e-05,
+      "loss": 0.7067,
+      "step": 1110
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.917036622903153e-05,
+      "loss": 0.5801,
+      "step": 1111
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.9118591826846003e-05,
+      "loss": 0.6165,
+      "step": 1112
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.906685471712096e-05,
+      "loss": 0.6422,
+      "step": 1113
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.9015155077168385e-05,
+      "loss": 0.6845,
+      "step": 1114
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.8963493084171842e-05,
+      "loss": 0.7465,
+      "step": 1115
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.8911868915185906e-05,
+      "loss": 0.6255,
+      "step": 1116
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.886028274713548e-05,
+      "loss": 0.7332,
+      "step": 1117
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.8808734756815253e-05,
+      "loss": 0.7103,
+      "step": 1118
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.8757225120889066e-05,
+      "loss": 0.6521,
+      "step": 1119
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.870575401588931e-05,
+      "loss": 0.8161,
+      "step": 1120
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.8654321618216344e-05,
+      "loss": 0.7023,
+      "step": 1121
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.8602928104137853e-05,
+      "loss": 0.6816,
+      "step": 1122
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.855157364978826e-05,
+      "loss": 0.7118,
+      "step": 1123
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.8500258431168127e-05,
+      "loss": 0.6067,
+      "step": 1124
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.8448982624143532e-05,
+      "loss": 0.675,
+      "step": 1125
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.8397746404445538e-05,
+      "loss": 0.657,
+      "step": 1126
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.834654994766947e-05,
+      "loss": 0.7183,
+      "step": 1127
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.829539342927441e-05,
+      "loss": 0.6412,
+      "step": 1128
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.8244277024582564e-05,
+      "loss": 0.8324,
+      "step": 1129
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.819320090877865e-05,
+      "loss": 0.6434,
+      "step": 1130
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.814216525690932e-05,
+      "loss": 0.6691,
+      "step": 1131
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.809117024388255e-05,
+      "loss": 0.6812,
+      "step": 1132
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.8040216044467032e-05,
+      "loss": 0.6041,
+      "step": 1133
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.798930283329159e-05,
+      "loss": 0.7071,
+      "step": 1134
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.793843078484456e-05,
+      "loss": 0.7947,
+      "step": 1135
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.7887600073473215e-05,
+      "loss": 0.7988,
+      "step": 1136
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.7836810873383172e-05,
+      "loss": 0.7014,
+      "step": 1137
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.7786063358637773e-05,
+      "loss": 0.7344,
+      "step": 1138
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.773535770315748e-05,
+      "loss": 0.8289,
+      "step": 1139
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.768469408071932e-05,
+      "loss": 0.6702,
+      "step": 1140
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.763407266495624e-05,
+      "loss": 0.6289,
+      "step": 1141
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.7583493629356572e-05,
+      "loss": 0.5965,
+      "step": 1142
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.753295714726338e-05,
+      "loss": 0.7443,
+      "step": 1143
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.7482463391873887e-05,
+      "loss": 0.769,
+      "step": 1144
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.7432012536238898e-05,
+      "loss": 0.7523,
+      "step": 1145
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.738160475326218e-05,
+      "loss": 0.8425,
+      "step": 1146
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.73312402156999e-05,
+      "loss": 0.8588,
+      "step": 1147
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.728091909616001e-05,
+      "loss": 0.6687,
+      "step": 1148
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.723064156710164e-05,
+      "loss": 0.6634,
+      "step": 1149
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.718040780083455e-05,
+      "loss": 0.6668,
+      "step": 1150
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.71302179695185e-05,
+      "loss": 0.8701,
+      "step": 1151
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.7080072245162715e-05,
+      "loss": 0.7142,
+      "step": 1152
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.7029970799625217e-05,
+      "loss": 0.6675,
+      "step": 1153
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.6979913804612293e-05,
+      "loss": 0.7825,
+      "step": 1154
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.6929901431677885e-05,
+      "loss": 0.72,
+      "step": 1155
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.6879933852223e-05,
+      "loss": 0.8381,
+      "step": 1156
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.6830011237495174e-05,
+      "loss": 0.6991,
+      "step": 1157
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.6780133758587782e-05,
+      "loss": 0.7356,
+      "step": 1158
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.673030158643955e-05,
+      "loss": 0.6712,
+      "step": 1159
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.6680514891833908e-05,
+      "loss": 0.6645,
+      "step": 1160
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.6630773845398427e-05,
+      "loss": 0.7014,
+      "step": 1161
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.6581078617604274e-05,
+      "loss": 0.7766,
+      "step": 1162
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.6531429378765536e-05,
+      "loss": 0.7153,
+      "step": 1163
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.6481826299038704e-05,
+      "loss": 0.7758,
+      "step": 1164
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.6432269548422087e-05,
+      "loss": 0.6753,
+      "step": 1165
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.6382759296755204e-05,
+      "loss": 0.7434,
+      "step": 1166
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.6333295713718206e-05,
+      "loss": 0.7652,
+      "step": 1167
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.6283878968831333e-05,
+      "loss": 0.6224,
+      "step": 1168
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.6234509231454277e-05,
+      "loss": 0.6707,
+      "step": 1169
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.6185186670785632e-05,
+      "loss": 0.752,
+      "step": 1170
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.613591145586231e-05,
+      "loss": 0.6293,
+      "step": 1171
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.6086683755558952e-05,
+      "loss": 0.6642,
+      "step": 1172
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.603750373858739e-05,
+      "loss": 0.7629,
+      "step": 1173
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.5988371573495998e-05,
+      "loss": 0.7175,
+      "step": 1174
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.5939287428669168e-05,
+      "loss": 0.743,
+      "step": 1175
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.589025147232674e-05,
+      "loss": 0.7385,
+      "step": 1176
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.5841263872523373e-05,
+      "loss": 0.7858,
+      "step": 1177
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.5792324797147996e-05,
+      "loss": 0.7718,
+      "step": 1178
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.574343441392328e-05,
+      "loss": 0.5513,
+      "step": 1179
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.5694592890404976e-05,
+      "loss": 0.7526,
+      "step": 1180
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.5645800393981402e-05,
+      "loss": 0.6316,
+      "step": 1181
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.5597057091872838e-05,
+      "loss": 0.7379,
+      "step": 1182
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.5548363151131e-05,
+      "loss": 0.6938,
+      "step": 1183
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.54997187386384e-05,
+      "loss": 0.7627,
+      "step": 1184
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.545112402110782e-05,
+      "loss": 0.7092,
+      "step": 1185
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.540257916508173e-05,
+      "loss": 0.6442,
+      "step": 1186
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.5354084336931707e-05,
+      "loss": 0.6888,
+      "step": 1187
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.5305639702857895e-05,
+      "loss": 0.6899,
+      "step": 1188
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.5257245428888397e-05,
+      "loss": 0.6168,
+      "step": 1189
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.5208901680878726e-05,
+      "loss": 0.7594,
+      "step": 1190
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.5160608624511238e-05,
+      "loss": 0.6841,
+      "step": 1191
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.5112366425294541e-05,
+      "loss": 0.7183,
+      "step": 1192
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.5064175248562988e-05,
+      "loss": 0.8455,
+      "step": 1193
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.5016035259476032e-05,
+      "loss": 0.7439,
+      "step": 1194
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.496794662301771e-05,
+      "loss": 0.638,
+      "step": 1195
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.4919909503996063e-05,
+      "loss": 0.5528,
+      "step": 1196
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.4871924067042563e-05,
+      "loss": 0.7507,
+      "step": 1197
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.4823990476611592e-05,
+      "loss": 0.7686,
+      "step": 1198
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4776108896979811e-05,
+      "loss": 0.7378,
+      "step": 1199
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4728279492245647e-05,
+      "loss": 0.7201,
+      "step": 1200
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4680502426328704e-05,
+      "loss": 0.7294,
+      "step": 1201
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4632777862969217e-05,
+      "loss": 0.7165,
+      "step": 1202
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4585105965727482e-05,
+      "loss": 0.781,
+      "step": 1203
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4537486897983323e-05,
+      "loss": 0.8344,
+      "step": 1204
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4489920822935475e-05,
+      "loss": 0.7604,
+      "step": 1205
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4442407903601079e-05,
+      "loss": 0.6557,
+      "step": 1206
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4394948302815088e-05,
+      "loss": 0.7388,
+      "step": 1207
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4347542183229721e-05,
+      "loss": 0.7839,
+      "step": 1208
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4300189707313945e-05,
+      "loss": 0.766,
+      "step": 1209
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4252891037352834e-05,
+      "loss": 0.7245,
+      "step": 1210
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4205646335447084e-05,
+      "loss": 0.6735,
+      "step": 1211
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4158455763512414e-05,
+      "loss": 0.746,
+      "step": 1212
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4111319483279063e-05,
+      "loss": 0.6507,
+      "step": 1213
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4064237656291183e-05,
+      "loss": 0.8152,
+      "step": 1214
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.4017210443906287e-05,
+      "loss": 0.7951,
+      "step": 1215
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.3970238007294757e-05,
+      "loss": 0.7911,
+      "step": 1216
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.3923320507439218e-05,
+      "loss": 0.7998,
+      "step": 1217
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.387645810513401e-05,
+      "loss": 0.9003,
+      "step": 1218
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.3829650960984685e-05,
+      "loss": 0.7251,
+      "step": 1219
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.3782899235407375e-05,
+      "loss": 0.8421,
+      "step": 1220
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.3736203088628304e-05,
+      "loss": 0.6894,
+      "step": 1221
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.3689562680683206e-05,
+      "loss": 0.804,
+      "step": 1222
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.3642978171416786e-05,
+      "loss": 0.8341,
+      "step": 1223
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.3596449720482203e-05,
+      "loss": 0.7345,
+      "step": 1224
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.3549977487340467e-05,
+      "loss": 0.8026,
+      "step": 1225
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.3503561631259924e-05,
+      "loss": 0.8868,
+      "step": 1226
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.3457202311315714e-05,
+      "loss": 0.9218,
+      "step": 1227
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.3410899686389194e-05,
+      "loss": 0.7833,
+      "step": 1228
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.3364653915167465e-05,
+      "loss": 0.732,
+      "step": 1229
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.3318465156142744e-05,
+      "loss": 0.708,
+      "step": 1230
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.3272333567611857e-05,
+      "loss": 0.7285,
+      "step": 1231
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.3226259307675711e-05,
+      "loss": 0.825,
+      "step": 1232
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.3180242534238723e-05,
+      "loss": 0.8444,
+      "step": 1233
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.313428340500832e-05,
+      "loss": 0.7145,
+      "step": 1234
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.308838207749435e-05,
+      "loss": 0.675,
+      "step": 1235
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.3042538709008568e-05,
+      "loss": 0.7996,
+      "step": 1236
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.2996753456664097e-05,
+      "loss": 0.7077,
+      "step": 1237
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.2951026477374882e-05,
+      "loss": 0.6692,
+      "step": 1238
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.2905357927855153e-05,
+      "loss": 0.8044,
+      "step": 1239
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.2859747964618915e-05,
+      "loss": 0.7766,
+      "step": 1240
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.2814196743979355e-05,
+      "loss": 0.846,
+      "step": 1241
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.2768704422048365e-05,
+      "loss": 0.7149,
+      "step": 1242
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.272327115473596e-05,
+      "loss": 0.8065,
+      "step": 1243
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.2677897097749768e-05,
+      "loss": 0.837,
+      "step": 1244
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.263258240659452e-05,
+      "loss": 0.7389,
+      "step": 1245
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.2587327236571457e-05,
+      "loss": 0.8551,
+      "step": 1246
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.2542131742777845e-05,
+      "loss": 0.9087,
+      "step": 1247
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.249699608010642e-05,
+      "loss": 0.8362,
+      "step": 1248
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.2451920403244888e-05,
+      "loss": 0.8026,
+      "step": 1249
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.2406904866675354e-05,
+      "loss": 0.824,
+      "step": 1250
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.2361949624673807e-05,
+      "loss": 0.8129,
+      "step": 1251
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.2317054831309626e-05,
+      "loss": 0.7929,
+      "step": 1252
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.2272220640444992e-05,
+      "loss": 0.8568,
+      "step": 1253
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.22274472057344e-05,
+      "loss": 0.6599,
+      "step": 1254
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.2182734680624115e-05,
+      "loss": 0.6863,
+      "step": 1255
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.2138083218351685e-05,
+      "loss": 0.9985,
+      "step": 1256
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.2093492971945355e-05,
+      "loss": 0.8038,
+      "step": 1257
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.2048964094223577e-05,
+      "loss": 0.8079,
+      "step": 1258
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.2004496737794478e-05,
+      "loss": 0.6987,
+      "step": 1259
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.1960091055055374e-05,
+      "loss": 0.7886,
+      "step": 1260
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.1915747198192176e-05,
+      "loss": 0.8175,
+      "step": 1261
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.187146531917892e-05,
+      "loss": 0.741,
+      "step": 1262
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.1827245569777236e-05,
+      "loss": 0.7373,
+      "step": 1263
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.1783088101535814e-05,
+      "loss": 0.751,
+      "step": 1264
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.1738993065789923e-05,
+      "loss": 0.8152,
+      "step": 1265
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.1694960613660842e-05,
+      "loss": 0.8015,
+      "step": 1266
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.1650990896055369e-05,
+      "loss": 0.8395,
+      "step": 1267
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.1607084063665292e-05,
+      "loss": 0.7202,
+      "step": 1268
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.1563240266966895e-05,
+      "loss": 0.9587,
+      "step": 1269
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.1519459656220434e-05,
+      "loss": 0.8308,
+      "step": 1270
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.1475742381469595e-05,
+      "loss": 0.9012,
+      "step": 1271
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.143208859254101e-05,
+      "loss": 0.7489,
+      "step": 1272
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.138849843904374e-05,
+      "loss": 0.8647,
+      "step": 1273
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.1344972070368745e-05,
+      "loss": 0.8657,
+      "step": 1274
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.1301509635688379e-05,
+      "loss": 0.8042,
+      "step": 1275
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.1258111283955917e-05,
+      "loss": 0.8915,
+      "step": 1276
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.1214777163904979e-05,
+      "loss": 0.9336,
+      "step": 1277
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.1171507424049051e-05,
+      "loss": 0.8286,
+      "step": 1278
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.1128302212680995e-05,
+      "loss": 0.918,
+      "step": 1279
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.1085161677872499e-05,
+      "loss": 0.8101,
+      "step": 1280
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.104208596747363e-05,
+      "loss": 0.7847,
+      "step": 1281
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.0999075229112257e-05,
+      "loss": 0.8828,
+      "step": 1282
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.0956129610193597e-05,
+      "loss": 0.8338,
+      "step": 1283
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.0913249257899674e-05,
+      "loss": 0.837,
+      "step": 1284
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.0870434319188842e-05,
+      "loss": 0.6845,
+      "step": 1285
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.0827684940795289e-05,
+      "loss": 0.7968,
+      "step": 1286
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.0785001269228487e-05,
+      "loss": 0.7873,
+      "step": 1287
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.0742383450772724e-05,
+      "loss": 0.9024,
+      "step": 1288
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.069983163148663e-05,
+      "loss": 0.909,
+      "step": 1289
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.0657345957202605e-05,
+      "loss": 0.7618,
+      "step": 1290
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.0614926573526366e-05,
+      "loss": 0.7524,
+      "step": 1291
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.0572573625836468e-05,
+      "loss": 0.8346,
+      "step": 1292
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.0530287259283752e-05,
+      "loss": 0.8522,
+      "step": 1293
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.048806761879088e-05,
+      "loss": 0.7612,
+      "step": 1294
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.0445914849051826e-05,
+      "loss": 0.8199,
+      "step": 1295
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.0403829094531409e-05,
+      "loss": 0.8129,
+      "step": 1296
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.0361810499464755e-05,
+      "loss": 0.8536,
+      "step": 1297
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.0319859207856826e-05,
+      "loss": 0.7941,
+      "step": 1298
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.0277975363481928e-05,
+      "loss": 0.8237,
+      "step": 1299
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.0236159109883197e-05,
+      "loss": 0.8334,
+      "step": 1300
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.0194410590372165e-05,
+      "loss": 0.905,
+      "step": 1301
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.0152729948028183e-05,
+      "loss": 0.7936,
+      "step": 1302
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.0111117325697999e-05,
+      "loss": 0.7619,
+      "step": 1303
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.0069572865995238e-05,
+      "loss": 0.8751,
+      "step": 1304
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.0028096711299907e-05,
+      "loss": 0.8348,
+      "step": 1305
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 9.986689003757957e-06,
+      "loss": 0.789,
+      "step": 1306
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 9.945349885280725e-06,
+      "loss": 0.7916,
+      "step": 1307
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 9.904079497544497e-06,
+      "loss": 0.8478,
+      "step": 1308
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 9.86287798198999e-06,
+      "loss": 0.7049,
+      "step": 1309
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 9.821745479821907e-06,
+      "loss": 0.7857,
+      "step": 1310
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 9.780682132008404e-06,
+      "loss": 0.6976,
+      "step": 1311
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 9.73968807928067e-06,
+      "loss": 0.7828,
+      "step": 1312
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 9.698763462132379e-06,
+      "loss": 0.9238,
+      "step": 1313
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 9.657908420819238e-06,
+      "loss": 0.8616,
+      "step": 1314
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 9.617123095358516e-06,
+      "loss": 0.804,
+      "step": 1315
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 9.576407625528542e-06,
+      "loss": 0.6784,
+      "step": 1316
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 9.535762150868263e-06,
+      "loss": 0.7249,
+      "step": 1317
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 9.495186810676711e-06,
+      "loss": 0.8955,
+      "step": 1318
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 9.454681744012571e-06,
+      "loss": 0.8126,
+      "step": 1319
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 9.414247089693681e-06,
+      "loss": 0.9375,
+      "step": 1320
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 9.373882986296564e-06,
+      "loss": 0.7388,
+      "step": 1321
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 9.33358957215597e-06,
+      "loss": 0.8288,
+      "step": 1322
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 9.293366985364366e-06,
+      "loss": 0.7996,
+      "step": 1323
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 9.253215363771483e-06,
+      "loss": 0.8633,
+      "step": 1324
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 9.213134844983834e-06,
+      "loss": 0.8469,
+      "step": 1325
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 9.17312556636429e-06,
+      "loss": 0.8105,
+      "step": 1326
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 9.133187665031522e-06,
+      "loss": 0.8902,
+      "step": 1327
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 9.093321277859594e-06,
+      "loss": 0.8215,
+      "step": 1328
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 9.0535265414775e-06,
+      "loss": 0.9031,
+      "step": 1329
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 9.013803592268652e-06,
+      "loss": 0.8164,
+      "step": 1330
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.974152566370428e-06,
+      "loss": 0.9964,
+      "step": 1331
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.934573599673745e-06,
+      "loss": 0.8584,
+      "step": 1332
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.895066827822529e-06,
+      "loss": 0.8536,
+      "step": 1333
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.855632386213293e-06,
+      "loss": 0.7563,
+      "step": 1334
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.816270409994669e-06,
+      "loss": 0.7476,
+      "step": 1335
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.776981034066912e-06,
+      "loss": 0.9192,
+      "step": 1336
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.737764393081505e-06,
+      "loss": 0.8603,
+      "step": 1337
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.698620621440628e-06,
+      "loss": 0.8525,
+      "step": 1338
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.659549853296723e-06,
+      "loss": 0.7207,
+      "step": 1339
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.62055222255205e-06,
+      "loss": 0.8311,
+      "step": 1340
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.581627862858196e-06,
+      "loss": 0.8513,
+      "step": 1341
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.542776907615662e-06,
+      "loss": 0.8441,
+      "step": 1342
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.503999489973362e-06,
+      "loss": 1.0006,
+      "step": 1343
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.465295742828184e-06,
+      "loss": 0.7336,
+      "step": 1344
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.42666579882454e-06,
+      "loss": 0.8971,
+      "step": 1345
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.3881097903539e-06,
+      "loss": 0.9649,
+      "step": 1346
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.34962784955434e-06,
+      "loss": 0.8902,
+      "step": 1347
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.311220108310119e-06,
+      "loss": 0.9103,
+      "step": 1348
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.272886698251171e-06,
+      "loss": 0.7893,
+      "step": 1349
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.234627750752702e-06,
+      "loss": 0.7654,
+      "step": 1350
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.196443396934716e-06,
+      "loss": 0.6874,
+      "step": 1351
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.158333767661557e-06,
+      "loss": 0.7692,
+      "step": 1352
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.120298993541514e-06,
+      "loss": 0.9211,
+      "step": 1353
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.082339204926297e-06,
+      "loss": 0.7689,
+      "step": 1354
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.044454531910646e-06,
+      "loss": 0.9029,
+      "step": 1355
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.00664510433186e-06,
+      "loss": 1.0813,
+      "step": 1356
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 7.96891105176935e-06,
+      "loss": 0.8261,
+      "step": 1357
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 7.93125250354424e-06,
+      "loss": 0.8527,
+      "step": 1358
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 7.893669588718844e-06,
+      "loss": 0.843,
+      "step": 1359
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 7.856162436096294e-06,
+      "loss": 0.8165,
+      "step": 1360
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 7.818731174220056e-06,
+      "loss": 0.842,
+      "step": 1361
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 7.78137593137352e-06,
+      "loss": 0.8933,
+      "step": 1362
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 7.744096835579539e-06,
+      "loss": 0.8831,
+      "step": 1363
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 7.706894014599994e-06,
+      "loss": 0.834,
+      "step": 1364
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 7.669767595935358e-06,
+      "loss": 0.8306,
+      "step": 1365
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 7.632717706824273e-06,
+      "loss": 0.9469,
+      "step": 1366
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 7.595744474243089e-06,
+      "loss": 0.874,
+      "step": 1367
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 7.558848024905435e-06,
+      "loss": 0.7326,
+      "step": 1368
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 7.522028485261814e-06,
+      "loss": 0.8573,
+      "step": 1369
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 7.4852859814991235e-06,
+      "loss": 1.002,
+      "step": 1370
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 7.448620639540256e-06,
+      "loss": 0.8139,
+      "step": 1371
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 7.41203258504365e-06,
+      "loss": 0.7543,
+      "step": 1372
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 7.375521943402879e-06,
+      "loss": 0.967,
+      "step": 1373
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 7.339088839746201e-06,
+      "loss": 0.9373,
+      "step": 1374
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 7.302733398936136e-06,
+      "loss": 0.8668,
+      "step": 1375
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 7.266455745569046e-06,
+      "loss": 0.8992,
+      "step": 1376
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 7.230256003974689e-06,
+      "loss": 0.9574,
+      "step": 1377
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 7.194134298215832e-06,
+      "loss": 0.8469,
+      "step": 1378
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 7.158090752087776e-06,
+      "loss": 0.8849,
+      "step": 1379
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 7.1221254891179625e-06,
+      "loss": 0.8809,
+      "step": 1380
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 7.086238632565544e-06,
+      "loss": 0.8484,
+      "step": 1381
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 7.050430305420961e-06,
+      "loss": 0.9786,
+      "step": 1382
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 7.014700630405513e-06,
+      "loss": 0.9801,
+      "step": 1383
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 6.979049729970969e-06,
+      "loss": 0.7151,
+      "step": 1384
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 6.9434777262990945e-06,
+      "loss": 0.8636,
+      "step": 1385
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 6.907984741301275e-06,
+      "loss": 0.9098,
+      "step": 1386
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 6.872570896618089e-06,
+      "loss": 0.8938,
+      "step": 1387
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 6.83723631361887e-06,
+      "loss": 0.8347,
+      "step": 1388
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 6.801981113401342e-06,
+      "loss": 0.8042,
+      "step": 1389
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 6.766805416791138e-06,
+      "loss": 0.9877,
+      "step": 1390
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 6.731709344341437e-06,
+      "loss": 0.9643,
+      "step": 1391
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.696693016332519e-06,
+      "loss": 0.79,
+      "step": 1392
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.661756552771366e-06,
+      "loss": 0.9065,
+      "step": 1393
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.626900073391276e-06,
+      "loss": 0.9984,
+      "step": 1394
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.592123697651399e-06,
+      "loss": 0.8666,
+      "step": 1395
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.557427544736365e-06,
+      "loss": 0.9297,
+      "step": 1396
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.5228117335558736e-06,
+      "loss": 0.8915,
+      "step": 1397
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.488276382744259e-06,
+      "loss": 0.9384,
+      "step": 1398
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.453821610660135e-06,
+      "loss": 0.9096,
+      "step": 1399
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.419447535385936e-06,
+      "loss": 0.9949,
+      "step": 1400
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.385154274727533e-06,
+      "loss": 0.8687,
+      "step": 1401
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.350941946213849e-06,
+      "loss": 0.7848,
+      "step": 1402
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.316810667096422e-06,
+      "loss": 0.968,
+      "step": 1403
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.282760554349021e-06,
+      "loss": 1.0317,
+      "step": 1404
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.248791724667264e-06,
+      "loss": 0.8551,
+      "step": 1405
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.214904294468172e-06,
+      "loss": 0.9323,
+      "step": 1406
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.1810983798898066e-06,
+      "loss": 0.8712,
+      "step": 1407
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 6.147374096790854e-06,
+      "loss": 0.8468,
+      "step": 1408
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 6.1137315607502575e-06,
+      "loss": 0.8039,
+      "step": 1409
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 6.080170887066781e-06,
+      "loss": 0.9687,
+      "step": 1410
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 6.046692190758628e-06,
+      "loss": 0.9545,
+      "step": 1411
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 6.013295586563064e-06,
+      "loss": 0.9408,
+      "step": 1412
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.9799811889359946e-06,
+      "loss": 0.9135,
+      "step": 1413
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.946749112051615e-06,
+      "loss": 0.9356,
+      "step": 1414
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.9135994698019734e-06,
+      "loss": 0.8994,
+      "step": 1415
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.880532375796602e-06,
+      "loss": 0.859,
+      "step": 1416
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.847547943362129e-06,
+      "loss": 0.9238,
+      "step": 1417
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.814646285541889e-06,
+      "loss": 0.8692,
+      "step": 1418
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.78182751509552e-06,
+      "loss": 0.8884,
+      "step": 1419
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.749091744498625e-06,
+      "loss": 0.8646,
+      "step": 1420
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.7164390859423125e-06,
+      "loss": 0.9275,
+      "step": 1421
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.683869651332877e-06,
+      "loss": 0.7904,
+      "step": 1422
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.651383552291378e-06,
+      "loss": 0.8595,
+      "step": 1423
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.618980900153267e-06,
+      "loss": 0.9843,
+      "step": 1424
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.586661805968029e-06,
+      "loss": 0.7902,
+      "step": 1425
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 5.554426380498763e-06,
+      "loss": 0.8104,
+      "step": 1426
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.522274734221824e-06,
+      "loss": 0.9032,
+      "step": 1427
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.490206977326443e-06,
+      "loss": 1.0298,
+      "step": 1428
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.458223219714338e-06,
+      "loss": 1.0443,
+      "step": 1429
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.426323570999377e-06,
+      "loss": 0.9272,
+      "step": 1430
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.394508140507137e-06,
+      "loss": 0.926,
+      "step": 1431
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.36277703727459e-06,
+      "loss": 0.9032,
+      "step": 1432
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.331130370049681e-06,
+      "loss": 1.0845,
+      "step": 1433
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.2995682472909945e-06,
+      "loss": 0.9362,
+      "step": 1434
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.268090777167369e-06,
+      "loss": 0.8547,
+      "step": 1435
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.236698067557515e-06,
+      "loss": 0.9306,
+      "step": 1436
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.2053902260496565e-06,
+      "loss": 0.8111,
+      "step": 1437
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.174167359941149e-06,
+      "loss": 0.8812,
+      "step": 1438
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.1430295762381554e-06,
+      "loss": 0.9364,
+      "step": 1439
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.111976981655209e-06,
+      "loss": 0.8839,
+      "step": 1440
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.0810096826149e-06,
+      "loss": 0.9499,
+      "step": 1441
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.05012778524751e-06,
+      "loss": 1.0657,
+      "step": 1442
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 5.019331395390616e-06,
+      "loss": 0.9362,
+      "step": 1443
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.988620618588744e-06,
+      "loss": 0.8718,
+      "step": 1444
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.9579955600930286e-06,
+      "loss": 1.041,
+      "step": 1445
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.927456324860819e-06,
+      "loss": 0.81,
+      "step": 1446
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.897003017555329e-06,
+      "loss": 0.8928,
+      "step": 1447
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.866635742545295e-06,
+      "loss": 0.9693,
+      "step": 1448
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.83635460390459e-06,
+      "loss": 0.8696,
+      "step": 1449
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.806159705411912e-06,
+      "loss": 0.8817,
+      "step": 1450
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.7760511505503655e-06,
+      "loss": 0.9685,
+      "step": 1451
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.746029042507163e-06,
+      "loss": 0.8608,
+      "step": 1452
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.716093484173236e-06,
+      "loss": 0.9744,
+      "step": 1453
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.686244578142902e-06,
+      "loss": 0.9849,
+      "step": 1454
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.656482426713495e-06,
+      "loss": 0.7988,
+      "step": 1455
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.626807131885051e-06,
+      "loss": 0.9543,
+      "step": 1456
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.597218795359913e-06,
+      "loss": 0.8658,
+      "step": 1457
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.567717518542408e-06,
+      "loss": 0.7524,
+      "step": 1458
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.53830340253849e-06,
+      "loss": 0.8602,
+      "step": 1459
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.508976548155397e-06,
+      "loss": 0.8934,
+      "step": 1460
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.4797370559013285e-06,
+      "loss": 0.9736,
+      "step": 1461
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.450585025985049e-06,
+      "loss": 0.8984,
+      "step": 1462
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.421520558315593e-06,
+      "loss": 0.8597,
+      "step": 1463
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.392543752501894e-06,
+      "loss": 0.9021,
+      "step": 1464
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.363654707852454e-06,
+      "loss": 0.8885,
+      "step": 1465
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.3348535233750195e-06,
+      "loss": 1.0152,
+      "step": 1466
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.3061402977762025e-06,
+      "loss": 0.9569,
+      "step": 1467
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.277515129461176e-06,
+      "loss": 0.78,
+      "step": 1468
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.248978116533323e-06,
+      "loss": 1.0388,
+      "step": 1469
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.2205293567938976e-06,
+      "loss": 0.8989,
+      "step": 1470
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.1921689477417116e-06,
+      "loss": 0.978,
+      "step": 1471
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.163896986572766e-06,
+      "loss": 0.918,
+      "step": 1472
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.135713570179945e-06,
+      "loss": 0.94,
+      "step": 1473
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.1076187951526726e-06,
+      "loss": 0.9394,
+      "step": 1474
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.079612757776573e-06,
+      "loss": 0.993,
+      "step": 1475
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.051695554033177e-06,
+      "loss": 1.0475,
+      "step": 1476
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.02386727959954e-06,
+      "loss": 0.9638,
+      "step": 1477
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 3.996128029847957e-06,
+      "loss": 0.7293,
+      "step": 1478
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.9684778998456165e-06,
+      "loss": 1.0039,
+      "step": 1479
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.940916984354278e-06,
+      "loss": 0.8645,
+      "step": 1480
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.913445377829942e-06,
+      "loss": 0.8632,
+      "step": 1481
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.886063174422545e-06,
+      "loss": 0.77,
+      "step": 1482
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.858770467975614e-06,
+      "loss": 1.1204,
+      "step": 1483
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.8315673520259506e-06,
+      "loss": 0.8981,
+      "step": 1484
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.804453919803319e-06,
+      "loss": 0.9398,
+      "step": 1485
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.7774302642301284e-06,
+      "loss": 0.9535,
+      "step": 1486
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.750496477921105e-06,
+      "loss": 0.9546,
+      "step": 1487
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.723652653182964e-06,
+      "loss": 0.9234,
+      "step": 1488
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.696898882014128e-06,
+      "loss": 0.9112,
+      "step": 1489
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.6702352561043727e-06,
+      "loss": 1.0687,
+      "step": 1490
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.6436618668345344e-06,
+      "loss": 0.9719,
+      "step": 1491
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.61717880527621e-06,
+      "loss": 1.139,
+      "step": 1492
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.590786162191404e-06,
+      "loss": 0.9294,
+      "step": 1493
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.564484028032252e-06,
+      "loss": 1.0852,
+      "step": 1494
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.538272492940702e-06,
+      "loss": 0.8911,
+      "step": 1495
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.512151646748192e-06,
+      "loss": 1.139,
+      "step": 1496
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.4861215789753763e-06,
+      "loss": 1.0235,
+      "step": 1497
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.4601823788317755e-06,
+      "loss": 0.9474,
+      "step": 1498
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.4343341352154946e-06,
+      "loss": 0.8744,
+      "step": 1499
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.4085769367129185e-06,
+      "loss": 0.927,
+      "step": 1500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.3829108715984012e-06,
+      "loss": 0.8616,
+      "step": 1501
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.357336027833975e-06,
+      "loss": 1.0599,
+      "step": 1502
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.3318524930690363e-06,
+      "loss": 0.9735,
+      "step": 1503
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.3064603546400473e-06,
+      "loss": 0.8522,
+      "step": 1504
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.2811596995702396e-06,
+      "loss": 1.0396,
+      "step": 1505
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.255950614569313e-06,
+      "loss": 0.8843,
+      "step": 1506
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.230833186033154e-06,
+      "loss": 0.9871,
+      "step": 1507
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.2058075000435116e-06,
+      "loss": 0.8869,
+      "step": 1508
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.180873642367725e-06,
+      "loss": 1.0059,
+      "step": 1509
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.1560316984584235e-06,
+      "loss": 1.0564,
+      "step": 1510
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.1312817534532136e-06,
+      "loss": 0.9369,
+      "step": 1511
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.1066238921744404e-06,
+      "loss": 0.8909,
+      "step": 1512
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.0820581991288364e-06,
+      "loss": 0.9541,
+      "step": 1513
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 3.0575847585072526e-06,
+      "loss": 0.8739,
+      "step": 1514
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 3.033203654184402e-06,
+      "loss": 0.9771,
+      "step": 1515
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 3.0089149697185225e-06,
+      "loss": 0.9751,
+      "step": 1516
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.984718788351117e-06,
+      "loss": 0.8649,
+      "step": 1517
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.960615193006677e-06,
+      "loss": 0.8929,
+      "step": 1518
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.93660426629237e-06,
+      "loss": 1.0344,
+      "step": 1519
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.9126860904977803e-06,
+      "loss": 0.976,
+      "step": 1520
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.8888607475946082e-06,
+      "loss": 0.9904,
+      "step": 1521
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.865128319236424e-06,
+      "loss": 0.9393,
+      "step": 1522
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.84148888675834e-06,
+      "loss": 0.8336,
+      "step": 1523
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.8179425311767626e-06,
+      "loss": 1.0244,
+      "step": 1524
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.7944893331891143e-06,
+      "loss": 0.8859,
+      "step": 1525
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.7711293731735386e-06,
+      "loss": 0.9016,
+      "step": 1526
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.7478627311886427e-06,
+      "loss": 0.8795,
+      "step": 1527
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.7246894869732287e-06,
+      "loss": 0.9877,
+      "step": 1528
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.7016097199459845e-06,
+      "loss": 0.9788,
+      "step": 1529
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.678623509205255e-06,
+      "loss": 1.0181,
+      "step": 1530
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 2.6557309335287405e-06,
+      "loss": 0.9213,
+      "step": 1531
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.6329320713732364e-06,
+      "loss": 0.8061,
+      "step": 1532
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.6102270008743833e-06,
+      "loss": 1.0308,
+      "step": 1533
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.5876157998463666e-06,
+      "loss": 0.8907,
+      "step": 1534
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.5650985457816656e-06,
+      "loss": 1.0637,
+      "step": 1535
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.5426753158507863e-06,
+      "loss": 1.0893,
+      "step": 1536
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.520346186901997e-06,
+      "loss": 1.1327,
+      "step": 1537
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.4981112354610837e-06,
+      "loss": 0.9232,
+      "step": 1538
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.4759705377310494e-06,
+      "loss": 0.8933,
+      "step": 1539
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.4539241695918734e-06,
+      "loss": 1.0166,
+      "step": 1540
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.4319722066002668e-06,
+      "loss": 0.9613,
+      "step": 1541
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.410114723989376e-06,
+      "loss": 1.0946,
+      "step": 1542
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.388351796668573e-06,
+      "loss": 1.0227,
+      "step": 1543
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.366683499223152e-06,
+      "loss": 0.8726,
+      "step": 1544
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.345109905914108e-06,
+      "loss": 0.8709,
+      "step": 1545
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.3236310906778534e-06,
+      "loss": 0.9258,
+      "step": 1546
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.302247127125989e-06,
+      "loss": 1.2111,
+      "step": 1547
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 2.2809580885450486e-06,
+      "loss": 0.8475,
+      "step": 1548
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 2.259764047896232e-06,
+      "loss": 1.013,
+      "step": 1549
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 2.2386650778151673e-06,
+      "loss": 0.9316,
+      "step": 1550
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 2.217661250611659e-06,
+      "loss": 1.119,
+      "step": 1551
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 2.1967526382694427e-06,
+      "loss": 1.0654,
+      "step": 1552
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 2.175939312445938e-06,
+      "loss": 1.2025,
+      "step": 1553
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 2.155221344471987e-06,
+      "loss": 1.0319,
+      "step": 1554
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 2.1345988053516517e-06,
+      "loss": 0.9077,
+      "step": 1555
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 2.114071765761917e-06,
+      "loss": 0.8898,
+      "step": 1556
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 2.0936402960524815e-06,
+      "loss": 0.8909,
+      "step": 1557
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 2.07330446624552e-06,
+      "loss": 1.0633,
+      "step": 1558
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 2.0530643460354237e-06,
+      "loss": 0.87,
+      "step": 1559
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 2.032920004788564e-06,
+      "loss": 0.8301,
+      "step": 1560
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 2.0128715115430772e-06,
+      "loss": 1.1354,
+      "step": 1561
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.9929189350085976e-06,
+      "loss": 1.2357,
+      "step": 1562
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.9730623435660388e-06,
+      "loss": 0.8649,
+      "step": 1563
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.9533018052673724e-06,
+      "loss": 1.038,
+      "step": 1564
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.933637387835362e-06,
+      "loss": 0.9947,
+      "step": 1565
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.914069158663352e-06,
+      "loss": 0.7785,
+      "step": 1566
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.8945971848150367e-06,
+      "loss": 0.9306,
+      "step": 1567
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.8752215330242217e-06,
+      "loss": 1.0915,
+      "step": 1568
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.8559422696946128e-06,
+      "loss": 0.9897,
+      "step": 1569
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.8367594608995596e-06,
+      "loss": 0.9247,
+      "step": 1570
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.817673172381854e-06,
+      "loss": 1.0133,
+      "step": 1571
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.7986834695534948e-06,
+      "loss": 0.9878,
+      "step": 1572
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.7797904174954628e-06,
+      "loss": 0.9323,
+      "step": 1573
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.7609940809575065e-06,
+      "loss": 0.718,
+      "step": 1574
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.7422945243579102e-06,
+      "loss": 1.0493,
+      "step": 1575
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.7236918117832767e-06,
+      "loss": 1.0461,
+      "step": 1576
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.7051860069883007e-06,
+      "loss": 0.9287,
+      "step": 1577
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.6867771733955629e-06,
+      "loss": 0.8704,
+      "step": 1578
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.668465374095316e-06,
+      "loss": 0.9232,
+      "step": 1579
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.650250671845246e-06,
+      "loss": 0.8731,
+      "step": 1580
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.632133129070278e-06,
+      "loss": 1.0046,
+      "step": 1581
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.6141128078623468e-06,
+      "loss": 0.9236,
+      "step": 1582
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.5961897699802008e-06,
+      "loss": 0.9022,
+      "step": 1583
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.5783640768491814e-06,
+      "loss": 0.7719,
+      "step": 1584
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.56063578956101e-06,
+      "loss": 0.9711,
+      "step": 1585
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.5430049688735793e-06,
+      "loss": 1.0245,
+      "step": 1586
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.5254716752107478e-06,
+      "loss": 0.9485,
+      "step": 1587
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.5080359686621293e-06,
+      "loss": 0.9127,
+      "step": 1588
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.4906979089829076e-06,
+      "loss": 1.0554,
+      "step": 1589
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.473457555593588e-06,
+      "loss": 1.0078,
+      "step": 1590
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.4563149675798338e-06,
+      "loss": 0.8874,
+      "step": 1591
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.4392702036922556e-06,
+      "loss": 0.8655,
+      "step": 1592
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.4223233223461896e-06,
+      "loss": 1.0566,
+      "step": 1593
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.4054743816215132e-06,
+      "loss": 1.1055,
+      "step": 1594
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.3887234392624592e-06,
+      "loss": 1.0625,
+      "step": 1595
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.3720705526773859e-06,
+      "loss": 0.8154,
+      "step": 1596
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.3555157789386064e-06,
+      "loss": 0.9,
+      "step": 1597
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.3390591747821834e-06,
+      "loss": 0.9585,
+      "step": 1598
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.322700796607732e-06,
+      "loss": 0.9741,
+      "step": 1599
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.3064407004782364e-06,
+      "loss": 0.9552,
+      "step": 1600
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.2902789421198435e-06,
+      "loss": 1.0402,
+      "step": 1601
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.2742155769216868e-06,
+      "loss": 0.8775,
+      "step": 1602
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.2582506599356825e-06,
+      "loss": 0.9303,
+      "step": 1603
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.2423842458763501e-06,
+      "loss": 0.8325,
+      "step": 1604
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.2266163891206328e-06,
+      "loss": 0.8358,
+      "step": 1605
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.2109471437076869e-06,
+      "loss": 0.9606,
+      "step": 1606
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.195376563338716e-06,
+      "loss": 0.9149,
+      "step": 1607
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.179904701376784e-06,
+      "loss": 1.0615,
+      "step": 1608
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.1645316108466187e-06,
+      "loss": 1.1474,
+      "step": 1609
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.1492573444344589e-06,
+      "loss": 0.9585,
+      "step": 1610
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.1340819544878444e-06,
+      "loss": 0.8824,
+      "step": 1611
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.1190054930154525e-06,
+      "loss": 0.9307,
+      "step": 1612
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.1040280116869117e-06,
+      "loss": 0.9637,
+      "step": 1613
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.0891495618326285e-06,
+      "loss": 0.9864,
+      "step": 1614
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.0743701944436214e-06,
+      "loss": 0.9101,
+      "step": 1615
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.0596899601713228e-06,
+      "loss": 1.0858,
+      "step": 1616
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.045108909327428e-06,
+      "loss": 0.9822,
+      "step": 1617
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.0306270918837068e-06,
+      "loss": 0.9361,
+      "step": 1618
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.0162445574718415e-06,
+      "loss": 0.9047,
+      "step": 1619
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.001961355383263e-06,
+      "loss": 0.8678,
+      "step": 1620
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 9.877775345689577e-07,
+      "loss": 0.9598,
+      "step": 1621
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 9.73693143639328e-07,
+      "loss": 0.928,
+      "step": 1622
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 9.597082308640059e-07,
+      "loss": 0.9229,
+      "step": 1623
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 9.458228441716921e-07,
+      "loss": 0.9906,
+      "step": 1624
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 9.320370311500104e-07,
+      "loss": 1.0828,
+      "step": 1625
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 9.183508390453077e-07,
+      "loss": 1.1255,
+      "step": 1626
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 9.04764314762524e-07,
+      "loss": 1.1044,
+      "step": 1627
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 8.912775048650223e-07,
+      "loss": 1.0317,
+      "step": 1628
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 8.778904555744227e-07,
+      "loss": 0.9445,
+      "step": 1629
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 8.64603212770445e-07,
+      "loss": 1.0273,
+      "step": 1630
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 8.51415821990763e-07,
+      "loss": 0.8103,
+      "step": 1631
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 8.383283284308374e-07,
+      "loss": 1.0841,
+      "step": 1632
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 8.253407769437627e-07,
+      "loss": 1.0439,
+      "step": 1633
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 8.124532120401173e-07,
+      "loss": 0.9466,
+      "step": 1634
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 7.996656778877975e-07,
+      "loss": 1.0866,
+      "step": 1635
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 7.869782183119001e-07,
+      "loss": 0.9427,
+      "step": 1636
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 7.743908767945229e-07,
+      "loss": 1.0225,
+      "step": 1637
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 7.619036964746617e-07,
+      "loss": 0.9669,
+      "step": 1638
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 7.495167201480302e-07,
+      "loss": 0.853,
+      "step": 1639
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 7.372299902669266e-07,
+      "loss": 0.89,
+      "step": 1640
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 7.250435489401008e-07,
+      "loss": 1.0344,
+      "step": 1641
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 7.129574379325809e-07,
+      "loss": 1.1284,
+      "step": 1642
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 7.009716986655535e-07,
+      "loss": 0.9103,
+      "step": 1643
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 6.890863722162099e-07,
+      "loss": 1.0662,
+      "step": 1644
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 6.773014993176108e-07,
+      "loss": 1.0161,
+      "step": 1645
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 6.656171203585482e-07,
+      "loss": 0.8792,
+      "step": 1646
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 6.540332753834033e-07,
+      "loss": 1.0148,
+      "step": 1647
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 6.425500040920062e-07,
+      "loss": 0.998,
+      "step": 1648
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 6.311673458395062e-07,
+      "loss": 1.0602,
+      "step": 1649
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 6.198853396362314e-07,
+      "loss": 0.8445,
+      "step": 1650
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 6.087040241475694e-07,
+      "loss": 1.071,
+      "step": 1651
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 5.976234376938039e-07,
+      "loss": 0.922,
+      "step": 1652
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 5.866436182500212e-07,
+      "loss": 1.0867,
+      "step": 1653
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 5.757646034459508e-07,
+      "loss": 1.0145,
+      "step": 1654
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 5.649864305658515e-07,
+      "loss": 1.054,
+      "step": 1655
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 5.543091365483788e-07,
+      "loss": 0.9525,
+      "step": 1656
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 5.437327579864582e-07,
+      "loss": 1.1156,
+      "step": 1657
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 5.332573311271616e-07,
+      "loss": 1.1367,
+      "step": 1658
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 5.228828918715844e-07,
+      "loss": 1.0104,
+      "step": 1659
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 5.126094757747091e-07,
+      "loss": 0.9167,
+      "step": 1660
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 5.024371180453113e-07,
+      "loss": 1.078,
+      "step": 1661
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.923658535458108e-07,
+      "loss": 0.9365,
+      "step": 1662
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.823957167921644e-07,
+      "loss": 0.9865,
+      "step": 1663
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.72526741953746e-07,
+      "loss": 0.9573,
+      "step": 1664
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.6275896285324026e-07,
+      "loss": 1.0292,
+      "step": 1665
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.530924129665026e-07,
+      "loss": 1.0602,
+      "step": 1666
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.4352712542246933e-07,
+      "loss": 0.929,
+      "step": 1667
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.3406313300302427e-07,
+      "loss": 0.7548,
+      "step": 1668
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.247004681429056e-07,
+      "loss": 0.9104,
+      "step": 1669
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.1543916292957286e-07,
+      "loss": 1.0244,
+      "step": 1670
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.062792491031164e-07,
+      "loss": 0.8475,
+      "step": 1671
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.972207580561382e-07,
+      "loss": 0.7713,
+      "step": 1672
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.882637208336548e-07,
+      "loss": 1.0632,
+      "step": 1673
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.794081681329675e-07,
+      "loss": 1.0121,
+      "step": 1674
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.7065413030358264e-07,
+      "loss": 0.9502,
+      "step": 1675
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.6200163734709467e-07,
+      "loss": 0.9141,
+      "step": 1676
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.534507189170966e-07,
+      "loss": 0.9678,
+      "step": 1677
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.450014043190497e-07,
+      "loss": 1.0407,
+      "step": 1678
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.36653722510214e-07,
+      "loss": 0.9427,
+      "step": 1679
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.284077020995313e-07,
+      "loss": 0.8786,
+      "step": 1680
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.202633713475356e-07,
+      "loss": 1.0309,
+      "step": 1681
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.1222075816625284e-07,
+      "loss": 0.9631,
+      "step": 1682
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.0427989011910463e-07,
+      "loss": 1.0276,
+      "step": 1683
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 2.964407944208114e-07,
+      "loss": 0.899,
+      "step": 1684
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 2.8870349793730264e-07,
+      "loss": 1.0218,
+      "step": 1685
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 2.810680271856303e-07,
+      "loss": 0.9599,
+      "step": 1686
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 2.7353440833386533e-07,
+      "loss": 0.9091,
+      "step": 1687
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 2.6610266720102137e-07,
+      "loss": 0.9818,
+      "step": 1688
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 2.5877282925695467e-07,
+      "loss": 1.0235,
+      "step": 1689
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 2.515449196222808e-07,
+      "loss": 1.0361,
+      "step": 1690
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 2.44418963068288e-07,
+      "loss": 0.7805,
+      "step": 1691
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 2.3739498401686433e-07,
+      "loss": 1.0539,
+      "step": 1692
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 2.304730065403904e-07,
+      "loss": 0.9059,
+      "step": 1693
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 2.2365305436167993e-07,
+      "loss": 0.9032,
+      "step": 1694
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 2.1693515085387972e-07,
+      "loss": 1.0041,
+      "step": 1695
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 2.103193190404029e-07,
+      "loss": 0.9427,
+      "step": 1696
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 2.038055815948492e-07,
+      "loss": 0.9322,
+      "step": 1697
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.973939608409181e-07,
+      "loss": 1.0027,
+      "step": 1698
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.9108447875233247e-07,
+      "loss": 1.2078,
+      "step": 1699
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.8487715695278184e-07,
+      "loss": 0.9801,
+      "step": 1700
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.7877201671582245e-07,
+      "loss": 1.0256,
+      "step": 1701
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.7276907896482396e-07,
+      "loss": 0.9697,
+      "step": 1702
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.6686836427288964e-07,
+      "loss": 1.115,
+      "step": 1703
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.6106989286278295e-07,
+      "loss": 0.9578,
+      "step": 1704
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.553736846068643e-07,
+      "loss": 1.0405,
+      "step": 1705
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.4977975902702113e-07,
+      "loss": 1.0413,
+      "step": 1706
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.4428813529460128e-07,
+      "loss": 1.0632,
+      "step": 1707
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.3889883223034304e-07,
+      "loss": 0.9063,
+      "step": 1708
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.336118683043186e-07,
+      "loss": 0.9717,
+      "step": 1709
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.2842726163586394e-07,
+      "loss": 0.9711,
+      "step": 1710
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.2334502999352237e-07,
+      "loss": 1.0093,
+      "step": 1711
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.1836519079496789e-07,
+      "loss": 1.1673,
+      "step": 1712
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.1348776110697844e-07,
+      "loss": 0.995,
+      "step": 1713
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.0871275764533617e-07,
+      "loss": 1.0537,
+      "step": 1714
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.0404019677480725e-07,
+      "loss": 1.0663,
+      "step": 1715
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 9.947009450906208e-08,
+      "loss": 1.0059,
+      "step": 1716
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 9.500246651062528e-08,
+      "loss": 1.1313,
+      "step": 1717
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 9.063732809083236e-08,
+      "loss": 0.9679,
+      "step": 1718
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 8.637469420975985e-08,
+      "loss": 1.0586,
+      "step": 1719
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 8.221457947619194e-08,
+      "loss": 1.0503,
+      "step": 1720
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 7.815699814755384e-08,
+      "loss": 0.9837,
+      "step": 1721
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 7.42019641298819e-08,
+      "loss": 1.0781,
+      "step": 1722
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 7.034949097775356e-08,
+      "loss": 1.0086,
+      "step": 1723
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 6.659959189426413e-08,
+      "loss": 1.0376,
+      "step": 1724
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 6.295227973096008e-08,
+      "loss": 1.0737,
+      "step": 1725
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 5.9407566987805805e-08,
+      "loss": 0.9096,
+      "step": 1726
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 5.596546581314699e-08,
+      "loss": 0.8752,
+      "step": 1727
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 5.262598800365726e-08,
+      "loss": 1.2397,
+      "step": 1728
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.938914500430158e-08,
+      "loss": 0.9294,
+      "step": 1729
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.625494790829632e-08,
+      "loss": 1.0646,
+      "step": 1730
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.322340745707587e-08,
+      "loss": 0.9244,
+      "step": 1731
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.029453404025274e-08,
+      "loss": 0.9754,
+      "step": 1732
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.746833769558422e-08,
+      "loss": 0.8449,
+      "step": 1733
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.4744828108925766e-08,
+      "loss": 0.95,
+      "step": 1734
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.212401461421432e-08,
+      "loss": 1.0688,
+      "step": 1735
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.960590619343506e-08,
+      "loss": 0.9349,
+      "step": 1736
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.7190511476574698e-08,
+      "loss": 0.8527,
+      "step": 1737
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.4877838741608207e-08,
+      "loss": 0.8717,
+      "step": 1738
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.2667895914462168e-08,
+      "loss": 1.1024,
+      "step": 1739
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.056069056899146e-08,
+      "loss": 1.0975,
+      "step": 1740
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.8556229926955937e-08,
+      "loss": 1.1622,
+      "step": 1741
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.66545208579838e-08,
+      "loss": 1.0796,
+      "step": 1742
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.4855569879558274e-08,
+      "loss": 1.0416,
+      "step": 1743
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.3159383156997607e-08,
+      "loss": 0.985,
+      "step": 1744
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.1565966503425118e-08,
+      "loss": 0.9701,
+      "step": 1745
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.0075325379755862e-08,
+      "loss": 1.0477,
+      "step": 1746
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 8.687464894669984e-09,
+      "loss": 1.2079,
+      "step": 1747
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 7.402389804606058e-09,
+      "loss": 1.0762,
+      "step": 1748
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 6.220104513737779e-09,
+      "loss": 1.0534,
+      "step": 1749
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 5.140613073960632e-09,
+      "loss": 0.8973,
+      "step": 1750
+    },
+    {
+      "epoch": 1.0,
+      "step": 1750,
+      "total_flos": 2.2456725560465818e+17,
+      "train_loss": 0.7509448079211372,
+      "train_runtime": 20535.3694,
+      "train_samples_per_second": 1.364,
+      "train_steps_per_second": 0.085
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 1750,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 50000,
+  "total_flos": 2.2456725560465818e+17,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}