diff --git a/.gitattributes b/.gitattributes
index 79281f5320e1b5f3e2102893cf10358bc788d823..581fd0b47aaa654231ad67bf0d6021398bcd8418 100644
--- a/.gitattributes
+++ b/.gitattributes
@@ -6126,3 +6126,4 @@ Meta-Llama-3-8B-Instruct_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d
 Meta-Llama-3-8B-Instruct_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-0.5-num-63987-sd-1/checkpoint-61896/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 Meta-Llama-3-8B-Instruct_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-0.5-num-63987-sd-1/checkpoint-7737/tokenizer.json filter=lfs diff=lfs merge=lfs -text
 Meta-Llama-3-8B-Instruct_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-0.5-num-63987-sd-1/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/tokenizer.json filter=lfs diff=lfs merge=lfs -text
diff --git a/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/README.md b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..503a34a03e25483aa99213835fd87bfc8289a3fe
--- /dev/null
+++ b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/README.md
@@ -0,0 +1,202 @@
+---
+base_model: google/gemma-2-9b-it
+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 0.13.1
\ No newline at end of file
diff --git a/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/adapter_config.json b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..e04e4441e0c1b29f69c16b26c142944e440b8076
--- /dev/null
+++ b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/adapter_config.json
@@ -0,0 +1,29 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "google/gemma-2-9b-it",
+  "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": 32,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 64,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "q_proj",
+    "v_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}
\ No newline at end of file
diff --git a/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/adapter_model.safetensors b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/adapter_model.safetensors
new file mode 100644
index 0000000000000000000000000000000000000000..065df76beeda65cc8491975fec6c64dd3f06d0d0
--- /dev/null
+++ b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/adapter_model.safetensors
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4ce8cb66e633086937bb05740fe402ff6f354a42a289c3e622ec9962c9635518
+size 143153376
diff --git a/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/optimizer.pt b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/optimizer.pt
new file mode 100644
index 0000000000000000000000000000000000000000..cd262e5c7ec5877477ad0f41ed57eab86200fd35
--- /dev/null
+++ b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/optimizer.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0d20de08f145be75ce39b608265eae9bf482f5cb1153ec08a93d12df4e26db2e
+size 72886650
diff --git a/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/rng_state.pth b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/rng_state.pth
new file mode 100644
index 0000000000000000000000000000000000000000..c4d8335ba08df4c241e490d04a6d08b8bab497f8
--- /dev/null
+++ b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/rng_state.pth
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c58304ec5a15a645bb6b486370c0ffec8bf48e6cbd63951b0994a6c8b2733e1d
+size 14244
diff --git a/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/scheduler.pt b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/scheduler.pt
new file mode 100644
index 0000000000000000000000000000000000000000..3e74997fdf0af6da45d90796dc099ed122f3a94e
--- /dev/null
+++ b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/scheduler.pt
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f752719d085c8860528c71d6c6b05296b0a818986a455e29d557b40ee05ef810
+size 1064
diff --git a/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/special_tokens_map.json b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/special_tokens_map.json
new file mode 100644
index 0000000000000000000000000000000000000000..0acb52c84d6ea33178bee426ec6706bfba8ba637
--- /dev/null
+++ b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/special_tokens_map.json
@@ -0,0 +1,28 @@
+{
+  "additional_special_tokens": [
+    "<start_of_turn>",
+    "<end_of_turn>"
+  ],
+  "bos_token": {
+    "content": "<bos>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<eos>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<eos>",
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}
diff --git a/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/tokenizer.json b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/tokenizer.json
new file mode 100644
index 0000000000000000000000000000000000000000..a4a305d1de4d8f47c0252b4d7fe65a10dd8e2c22
--- /dev/null
+++ b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/tokenizer.json
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5f7eee611703c5ce5d1eee32d9cdcfe465647b8aff0c1dfb3bed7ad7dbb05060
+size 34362873
diff --git a/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/tokenizer.model b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/tokenizer.model
new file mode 100644
index 0000000000000000000000000000000000000000..796efe9ab515c15e146ce7588e6d7b9b8134dbf8
--- /dev/null
+++ b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/tokenizer.model
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:61a7b147390c64585d6c3543dd6fc636906c9af3865a5548f27f31aee1d4c8e2
+size 4241003
diff --git a/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/tokenizer_config.json b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..1aa249f4dc9f84e87ad8983458e7800ae5bf5454
--- /dev/null
+++ b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/tokenizer_config.json
@@ -0,0 +1,2013 @@
+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<pad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<eos>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "<bos>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "3": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "4": {
+      "content": "<mask>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "5": {
+      "content": "<2mass>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "6": {
+      "content": "[@BOS@]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "7": {
+      "content": "<unused0>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "8": {
+      "content": "<unused1>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "9": {
+      "content": "<unused2>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "10": {
+      "content": "<unused3>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "11": {
+      "content": "<unused4>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "12": {
+      "content": "<unused5>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "13": {
+      "content": "<unused6>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "14": {
+      "content": "<unused7>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "15": {
+      "content": "<unused8>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "16": {
+      "content": "<unused9>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "17": {
+      "content": "<unused10>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "18": {
+      "content": "<unused11>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "19": {
+      "content": "<unused12>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "20": {
+      "content": "<unused13>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "21": {
+      "content": "<unused14>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "22": {
+      "content": "<unused15>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "23": {
+      "content": "<unused16>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "24": {
+      "content": "<unused17>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "25": {
+      "content": "<unused18>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "26": {
+      "content": "<unused19>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "27": {
+      "content": "<unused20>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "28": {
+      "content": "<unused21>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "29": {
+      "content": "<unused22>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "30": {
+      "content": "<unused23>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "31": {
+      "content": "<unused24>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "32": {
+      "content": "<unused25>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "33": {
+      "content": "<unused26>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "34": {
+      "content": "<unused27>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "35": {
+      "content": "<unused28>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "36": {
+      "content": "<unused29>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "37": {
+      "content": "<unused30>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "38": {
+      "content": "<unused31>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "39": {
+      "content": "<unused32>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "40": {
+      "content": "<unused33>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "41": {
+      "content": "<unused34>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "42": {
+      "content": "<unused35>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "43": {
+      "content": "<unused36>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "44": {
+      "content": "<unused37>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "45": {
+      "content": "<unused38>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "46": {
+      "content": "<unused39>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "47": {
+      "content": "<unused40>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "48": {
+      "content": "<unused41>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "49": {
+      "content": "<unused42>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "50": {
+      "content": "<unused43>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "51": {
+      "content": "<unused44>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "52": {
+      "content": "<unused45>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "53": {
+      "content": "<unused46>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "54": {
+      "content": "<unused47>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "55": {
+      "content": "<unused48>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "56": {
+      "content": "<unused49>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "57": {
+      "content": "<unused50>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "58": {
+      "content": "<unused51>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "59": {
+      "content": "<unused52>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "60": {
+      "content": "<unused53>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "61": {
+      "content": "<unused54>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "62": {
+      "content": "<unused55>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "63": {
+      "content": "<unused56>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "64": {
+      "content": "<unused57>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "65": {
+      "content": "<unused58>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "66": {
+      "content": "<unused59>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "67": {
+      "content": "<unused60>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "68": {
+      "content": "<unused61>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "69": {
+      "content": "<unused62>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "70": {
+      "content": "<unused63>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "71": {
+      "content": "<unused64>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "72": {
+      "content": "<unused65>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "73": {
+      "content": "<unused66>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "74": {
+      "content": "<unused67>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "75": {
+      "content": "<unused68>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "76": {
+      "content": "<unused69>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "77": {
+      "content": "<unused70>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "78": {
+      "content": "<unused71>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "79": {
+      "content": "<unused72>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "80": {
+      "content": "<unused73>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "81": {
+      "content": "<unused74>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "82": {
+      "content": "<unused75>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "83": {
+      "content": "<unused76>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "84": {
+      "content": "<unused77>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "85": {
+      "content": "<unused78>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "86": {
+      "content": "<unused79>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "87": {
+      "content": "<unused80>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "88": {
+      "content": "<unused81>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "89": {
+      "content": "<unused82>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "90": {
+      "content": "<unused83>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "91": {
+      "content": "<unused84>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "92": {
+      "content": "<unused85>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "93": {
+      "content": "<unused86>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "94": {
+      "content": "<unused87>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "95": {
+      "content": "<unused88>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "96": {
+      "content": "<unused89>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "97": {
+      "content": "<unused90>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "98": {
+      "content": "<unused91>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "99": {
+      "content": "<unused92>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "100": {
+      "content": "<unused93>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "101": {
+      "content": "<unused94>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "102": {
+      "content": "<unused95>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "103": {
+      "content": "<unused96>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "104": {
+      "content": "<unused97>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "105": {
+      "content": "<unused98>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "106": {
+      "content": "<start_of_turn>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "107": {
+      "content": "<end_of_turn>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "108": {
+      "content": "\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "109": {
+      "content": "\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "110": {
+      "content": "\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "111": {
+      "content": "\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "112": {
+      "content": "\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "113": {
+      "content": "\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "114": {
+      "content": "\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "115": {
+      "content": "\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "116": {
+      "content": "\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "117": {
+      "content": "\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "118": {
+      "content": "\n\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "119": {
+      "content": "\n\n\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "120": {
+      "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "121": {
+      "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "122": {
+      "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "123": {
+      "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "124": {
+      "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "125": {
+      "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "126": {
+      "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "127": {
+      "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "128": {
+      "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "129": {
+      "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "130": {
+      "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "131": {
+      "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "132": {
+      "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "133": {
+      "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "134": {
+      "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "135": {
+      "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "136": {
+      "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "137": {
+      "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "138": {
+      "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "139": {
+      "content": "▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "140": {
+      "content": "▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "141": {
+      "content": "▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "142": {
+      "content": "▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "143": {
+      "content": "▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "144": {
+      "content": "▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "145": {
+      "content": "▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "146": {
+      "content": "▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "147": {
+      "content": "▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "148": {
+      "content": "▁▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "149": {
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "150": {
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151": {
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "152": {
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "153": {
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "154": {
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "155": {
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "156": {
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "157": {
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "158": {
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "159": {
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "160": {
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "161": {
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "162": {
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "163": {
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "164": {
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "165": {
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "166": {
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "167": {
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "168": {
+      "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "169": {
+      "content": "<table>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "170": {
+      "content": "<caption>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "171": {
+      "content": "<thead>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "172": {
+      "content": "<tbody>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "173": {
+      "content": "<tfoot>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "174": {
+      "content": "<tr>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "175": {
+      "content": "<th>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "176": {
+      "content": "<td>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "177": {
+      "content": "</table>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "178": {
+      "content": "</caption>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "179": {
+      "content": "</thead>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "180": {
+      "content": "</tbody>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "181": {
+      "content": "</tfoot>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "182": {
+      "content": "</tr>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "183": {
+      "content": "</th>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "184": {
+      "content": "</td>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "185": {
+      "content": "<h1>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "186": {
+      "content": "<h2>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "187": {
+      "content": "<h3>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "188": {
+      "content": "<h4>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "189": {
+      "content": "<h5>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "190": {
+      "content": "<h6>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "191": {
+      "content": "<blockquote>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "192": {
+      "content": "</h1>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "193": {
+      "content": "</h2>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "194": {
+      "content": "</h3>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "195": {
+      "content": "</h4>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "196": {
+      "content": "</h5>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "197": {
+      "content": "</h6>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "198": {
+      "content": "</blockquote>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "199": {
+      "content": "<strong>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "200": {
+      "content": "<em>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "201": {
+      "content": "<b>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "202": {
+      "content": "<i>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "203": {
+      "content": "<u>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "204": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "205": {
+      "content": "<sub>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "206": {
+      "content": "<sup>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "207": {
+      "content": "<code>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "208": {
+      "content": "</strong>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "209": {
+      "content": "</em>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "210": {
+      "content": "</b>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "211": {
+      "content": "</i>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "212": {
+      "content": "</u>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "213": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "214": {
+      "content": "</sub>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "215": {
+      "content": "</sup>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "216": {
+      "content": "</code>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255968": {
+      "content": "[toxicity=0]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255969": {
+      "content": "\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255970": {
+      "content": "\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255971": {
+      "content": "\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255972": {
+      "content": "\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255973": {
+      "content": "\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255974": {
+      "content": "\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255975": {
+      "content": "\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255976": {
+      "content": "\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255977": {
+      "content": "\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255978": {
+      "content": "\t\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255979": {
+      "content": "\t\t\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255980": {
+      "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255981": {
+      "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255982": {
+      "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255983": {
+      "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255984": {
+      "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255985": {
+      "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255986": {
+      "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255987": {
+      "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255988": {
+      "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255989": {
+      "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255990": {
+      "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255991": {
+      "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255992": {
+      "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255993": {
+      "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255994": {
+      "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255995": {
+      "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255996": {
+      "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255997": {
+      "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255998": {
+      "content": "\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "255999": {
+      "content": "<unused99>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<start_of_turn>",
+    "<end_of_turn>"
+  ],
+  "bos_token": "<bos>",
+  "chat_template": "{{ bos_token }}{% if messages[0]['role'] == 'system' %}{{ raise_exception('System role not supported') }}{% endif %}{% for message in messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if (message['role'] == 'assistant') %}{% set role = 'model' %}{% else %}{% set role = message['role'] %}{% endif %}{{ '<start_of_turn>' + role + '\n' + message['content'] | trim + '<end_of_turn>\n' }}{% endfor %}{% if add_generation_prompt %}{{'<start_of_turn>model\n'}}{% endif %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<eos>",
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<eos>",
+  "sp_model_kwargs": {},
+  "spaces_between_special_tokens": false,
+  "tokenizer_class": "GemmaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": false
+}
diff --git a/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/trainer_state.json b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..7bb4b4c8ebf2efe24240c827d48cf9fc192e6375
--- /dev/null
+++ b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/trainer_state.json
@@ -0,0 +1,10702 @@
+{
+  "best_metric": 1.055190086364746,
+  "best_model_checkpoint": "outputs-001/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237",
+  "epoch": 0.9999671862182117,
+  "eval_steps": 10,
+  "global_step": 15237,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0006562756357670221,
+      "grad_norm": 1.0711857080459595,
+      "learning_rate": 0.0002,
+      "loss": 2.6059,
+      "step": 10
+    },
+    {
+      "epoch": 0.0013125512715340443,
+      "grad_norm": 0.935492992401123,
+      "learning_rate": 0.0002,
+      "loss": 1.7643,
+      "step": 20
+    },
+    {
+      "epoch": 0.0019688269073010667,
+      "grad_norm": 0.908809244632721,
+      "learning_rate": 0.0002,
+      "loss": 1.2573,
+      "step": 30
+    },
+    {
+      "epoch": 0.0026251025430680886,
+      "grad_norm": 0.3497907221317291,
+      "learning_rate": 0.0002,
+      "loss": 1.1011,
+      "step": 40
+    },
+    {
+      "epoch": 0.003281378178835111,
+      "grad_norm": 0.32830339670181274,
+      "learning_rate": 0.0002,
+      "loss": 1.0866,
+      "step": 50
+    },
+    {
+      "epoch": 0.003937653814602133,
+      "grad_norm": 0.2850394546985626,
+      "learning_rate": 0.0002,
+      "loss": 1.0104,
+      "step": 60
+    },
+    {
+      "epoch": 0.004593929450369155,
+      "grad_norm": 0.3804827928543091,
+      "learning_rate": 0.0002,
+      "loss": 1.082,
+      "step": 70
+    },
+    {
+      "epoch": 0.005250205086136177,
+      "grad_norm": 0.23506930470466614,
+      "learning_rate": 0.0002,
+      "loss": 1.052,
+      "step": 80
+    },
+    {
+      "epoch": 0.005906480721903199,
+      "grad_norm": 0.2644859552383423,
+      "learning_rate": 0.0002,
+      "loss": 1.0593,
+      "step": 90
+    },
+    {
+      "epoch": 0.006562756357670222,
+      "grad_norm": 0.36523646116256714,
+      "learning_rate": 0.0002,
+      "loss": 0.9978,
+      "step": 100
+    },
+    {
+      "epoch": 0.007219031993437244,
+      "grad_norm": 0.3195570111274719,
+      "learning_rate": 0.0002,
+      "loss": 0.9958,
+      "step": 110
+    },
+    {
+      "epoch": 0.007875307629204267,
+      "grad_norm": 0.2886694371700287,
+      "learning_rate": 0.0002,
+      "loss": 1.0041,
+      "step": 120
+    },
+    {
+      "epoch": 0.008531583264971289,
+      "grad_norm": 0.3913154900074005,
+      "learning_rate": 0.0002,
+      "loss": 1.0623,
+      "step": 130
+    },
+    {
+      "epoch": 0.00918785890073831,
+      "grad_norm": 0.3181937336921692,
+      "learning_rate": 0.0002,
+      "loss": 1.1128,
+      "step": 140
+    },
+    {
+      "epoch": 0.009844134536505332,
+      "grad_norm": 0.2629619538784027,
+      "learning_rate": 0.0002,
+      "loss": 0.9989,
+      "step": 150
+    },
+    {
+      "epoch": 0.010500410172272354,
+      "grad_norm": 0.30438563227653503,
+      "learning_rate": 0.0002,
+      "loss": 1.0226,
+      "step": 160
+    },
+    {
+      "epoch": 0.011156685808039376,
+      "grad_norm": 0.2876931130886078,
+      "learning_rate": 0.0002,
+      "loss": 1.0321,
+      "step": 170
+    },
+    {
+      "epoch": 0.011812961443806398,
+      "grad_norm": 0.29188141226768494,
+      "learning_rate": 0.0002,
+      "loss": 0.9824,
+      "step": 180
+    },
+    {
+      "epoch": 0.01246923707957342,
+      "grad_norm": 0.2645126283168793,
+      "learning_rate": 0.0002,
+      "loss": 0.9439,
+      "step": 190
+    },
+    {
+      "epoch": 0.013125512715340444,
+      "grad_norm": 0.26031428575515747,
+      "learning_rate": 0.0002,
+      "loss": 1.0273,
+      "step": 200
+    },
+    {
+      "epoch": 0.013781788351107466,
+      "grad_norm": 0.25812748074531555,
+      "learning_rate": 0.0002,
+      "loss": 1.0518,
+      "step": 210
+    },
+    {
+      "epoch": 0.014438063986874488,
+      "grad_norm": 0.24913132190704346,
+      "learning_rate": 0.0002,
+      "loss": 0.9502,
+      "step": 220
+    },
+    {
+      "epoch": 0.01509433962264151,
+      "grad_norm": 0.30332663655281067,
+      "learning_rate": 0.0002,
+      "loss": 1.0131,
+      "step": 230
+    },
+    {
+      "epoch": 0.015750615258408533,
+      "grad_norm": 0.25207284092903137,
+      "learning_rate": 0.0002,
+      "loss": 1.0459,
+      "step": 240
+    },
+    {
+      "epoch": 0.016406890894175553,
+      "grad_norm": 0.26384010910987854,
+      "learning_rate": 0.0002,
+      "loss": 0.9798,
+      "step": 250
+    },
+    {
+      "epoch": 0.017063166529942577,
+      "grad_norm": 0.28651612997055054,
+      "learning_rate": 0.0002,
+      "loss": 1.061,
+      "step": 260
+    },
+    {
+      "epoch": 0.017719442165709597,
+      "grad_norm": 0.2879799008369446,
+      "learning_rate": 0.0002,
+      "loss": 0.991,
+      "step": 270
+    },
+    {
+      "epoch": 0.01837571780147662,
+      "grad_norm": 0.28661131858825684,
+      "learning_rate": 0.0002,
+      "loss": 0.9643,
+      "step": 280
+    },
+    {
+      "epoch": 0.01903199343724364,
+      "grad_norm": 0.265348345041275,
+      "learning_rate": 0.0002,
+      "loss": 0.9903,
+      "step": 290
+    },
+    {
+      "epoch": 0.019688269073010665,
+      "grad_norm": 0.24186863005161285,
+      "learning_rate": 0.0002,
+      "loss": 1.011,
+      "step": 300
+    },
+    {
+      "epoch": 0.020344544708777685,
+      "grad_norm": 0.25996068120002747,
+      "learning_rate": 0.0002,
+      "loss": 1.0245,
+      "step": 310
+    },
+    {
+      "epoch": 0.02100082034454471,
+      "grad_norm": 0.29615098237991333,
+      "learning_rate": 0.0002,
+      "loss": 0.9987,
+      "step": 320
+    },
+    {
+      "epoch": 0.021657095980311732,
+      "grad_norm": 0.2429388016462326,
+      "learning_rate": 0.0002,
+      "loss": 1.04,
+      "step": 330
+    },
+    {
+      "epoch": 0.022313371616078752,
+      "grad_norm": 0.26882505416870117,
+      "learning_rate": 0.0002,
+      "loss": 0.9703,
+      "step": 340
+    },
+    {
+      "epoch": 0.022969647251845776,
+      "grad_norm": 0.283328652381897,
+      "learning_rate": 0.0002,
+      "loss": 0.9686,
+      "step": 350
+    },
+    {
+      "epoch": 0.023625922887612796,
+      "grad_norm": 0.3115910589694977,
+      "learning_rate": 0.0002,
+      "loss": 0.952,
+      "step": 360
+    },
+    {
+      "epoch": 0.02428219852337982,
+      "grad_norm": 0.27969497442245483,
+      "learning_rate": 0.0002,
+      "loss": 1.0112,
+      "step": 370
+    },
+    {
+      "epoch": 0.02493847415914684,
+      "grad_norm": 0.30471885204315186,
+      "learning_rate": 0.0002,
+      "loss": 1.0618,
+      "step": 380
+    },
+    {
+      "epoch": 0.025594749794913864,
+      "grad_norm": 0.3183926045894623,
+      "learning_rate": 0.0002,
+      "loss": 1.0189,
+      "step": 390
+    },
+    {
+      "epoch": 0.026251025430680888,
+      "grad_norm": 0.27311646938323975,
+      "learning_rate": 0.0002,
+      "loss": 1.0148,
+      "step": 400
+    },
+    {
+      "epoch": 0.026907301066447908,
+      "grad_norm": 0.29148945212364197,
+      "learning_rate": 0.0002,
+      "loss": 0.9672,
+      "step": 410
+    },
+    {
+      "epoch": 0.02756357670221493,
+      "grad_norm": 0.2386617809534073,
+      "learning_rate": 0.0002,
+      "loss": 0.9473,
+      "step": 420
+    },
+    {
+      "epoch": 0.02821985233798195,
+      "grad_norm": 0.2546529471874237,
+      "learning_rate": 0.0002,
+      "loss": 1.0722,
+      "step": 430
+    },
+    {
+      "epoch": 0.028876127973748975,
+      "grad_norm": 0.27932611107826233,
+      "learning_rate": 0.0002,
+      "loss": 1.0017,
+      "step": 440
+    },
+    {
+      "epoch": 0.029532403609515995,
+      "grad_norm": 0.31259334087371826,
+      "learning_rate": 0.0002,
+      "loss": 0.9988,
+      "step": 450
+    },
+    {
+      "epoch": 0.03018867924528302,
+      "grad_norm": 0.2675893008708954,
+      "learning_rate": 0.0002,
+      "loss": 1.0404,
+      "step": 460
+    },
+    {
+      "epoch": 0.03084495488105004,
+      "grad_norm": 0.24047039449214935,
+      "learning_rate": 0.0002,
+      "loss": 1.0018,
+      "step": 470
+    },
+    {
+      "epoch": 0.031501230516817066,
+      "grad_norm": 0.2637856900691986,
+      "learning_rate": 0.0002,
+      "loss": 1.0685,
+      "step": 480
+    },
+    {
+      "epoch": 0.03215750615258409,
+      "grad_norm": 0.3064589500427246,
+      "learning_rate": 0.0002,
+      "loss": 0.9879,
+      "step": 490
+    },
+    {
+      "epoch": 0.03281378178835111,
+      "grad_norm": 0.25345391035079956,
+      "learning_rate": 0.0002,
+      "loss": 0.9657,
+      "step": 500
+    },
+    {
+      "epoch": 0.03347005742411813,
+      "grad_norm": 0.3100789785385132,
+      "learning_rate": 0.0002,
+      "loss": 0.9518,
+      "step": 510
+    },
+    {
+      "epoch": 0.034126333059885154,
+      "grad_norm": 0.3312002420425415,
+      "learning_rate": 0.0002,
+      "loss": 1.0108,
+      "step": 520
+    },
+    {
+      "epoch": 0.034782608695652174,
+      "grad_norm": 0.23432421684265137,
+      "learning_rate": 0.0002,
+      "loss": 1.0157,
+      "step": 530
+    },
+    {
+      "epoch": 0.035438884331419195,
+      "grad_norm": 0.3079119622707367,
+      "learning_rate": 0.0002,
+      "loss": 1.0093,
+      "step": 540
+    },
+    {
+      "epoch": 0.03609515996718622,
+      "grad_norm": 0.2555035650730133,
+      "learning_rate": 0.0002,
+      "loss": 0.9534,
+      "step": 550
+    },
+    {
+      "epoch": 0.03675143560295324,
+      "grad_norm": 0.3530837893486023,
+      "learning_rate": 0.0002,
+      "loss": 0.9648,
+      "step": 560
+    },
+    {
+      "epoch": 0.03740771123872026,
+      "grad_norm": 0.26614823937416077,
+      "learning_rate": 0.0002,
+      "loss": 1.0432,
+      "step": 570
+    },
+    {
+      "epoch": 0.03806398687448728,
+      "grad_norm": 0.2618412971496582,
+      "learning_rate": 0.0002,
+      "loss": 1.0108,
+      "step": 580
+    },
+    {
+      "epoch": 0.03872026251025431,
+      "grad_norm": 0.26110127568244934,
+      "learning_rate": 0.0002,
+      "loss": 0.9834,
+      "step": 590
+    },
+    {
+      "epoch": 0.03937653814602133,
+      "grad_norm": 0.3050612211227417,
+      "learning_rate": 0.0002,
+      "loss": 0.9766,
+      "step": 600
+    },
+    {
+      "epoch": 0.04003281378178835,
+      "grad_norm": 0.27394089102745056,
+      "learning_rate": 0.0002,
+      "loss": 1.0195,
+      "step": 610
+    },
+    {
+      "epoch": 0.04068908941755537,
+      "grad_norm": 0.25288277864456177,
+      "learning_rate": 0.0002,
+      "loss": 0.9776,
+      "step": 620
+    },
+    {
+      "epoch": 0.0413453650533224,
+      "grad_norm": 0.2835882902145386,
+      "learning_rate": 0.0002,
+      "loss": 0.9199,
+      "step": 630
+    },
+    {
+      "epoch": 0.04200164068908942,
+      "grad_norm": 0.2621902823448181,
+      "learning_rate": 0.0002,
+      "loss": 0.9605,
+      "step": 640
+    },
+    {
+      "epoch": 0.04265791632485644,
+      "grad_norm": 0.23218439519405365,
+      "learning_rate": 0.0002,
+      "loss": 0.9494,
+      "step": 650
+    },
+    {
+      "epoch": 0.043314191960623465,
+      "grad_norm": 0.26911118626594543,
+      "learning_rate": 0.0002,
+      "loss": 0.994,
+      "step": 660
+    },
+    {
+      "epoch": 0.043970467596390485,
+      "grad_norm": 0.25920751690864563,
+      "learning_rate": 0.0002,
+      "loss": 0.972,
+      "step": 670
+    },
+    {
+      "epoch": 0.044626743232157505,
+      "grad_norm": 0.2772065997123718,
+      "learning_rate": 0.0002,
+      "loss": 0.9774,
+      "step": 680
+    },
+    {
+      "epoch": 0.045283018867924525,
+      "grad_norm": 0.275421142578125,
+      "learning_rate": 0.0002,
+      "loss": 0.9114,
+      "step": 690
+    },
+    {
+      "epoch": 0.04593929450369155,
+      "grad_norm": 0.23931777477264404,
+      "learning_rate": 0.0002,
+      "loss": 1.028,
+      "step": 700
+    },
+    {
+      "epoch": 0.04659557013945857,
+      "grad_norm": 0.3031066060066223,
+      "learning_rate": 0.0002,
+      "loss": 0.9309,
+      "step": 710
+    },
+    {
+      "epoch": 0.04725184577522559,
+      "grad_norm": 0.2655068039894104,
+      "learning_rate": 0.0002,
+      "loss": 0.9511,
+      "step": 720
+    },
+    {
+      "epoch": 0.04790812141099262,
+      "grad_norm": 0.26064610481262207,
+      "learning_rate": 0.0002,
+      "loss": 0.9012,
+      "step": 730
+    },
+    {
+      "epoch": 0.04856439704675964,
+      "grad_norm": 0.2934698760509491,
+      "learning_rate": 0.0002,
+      "loss": 0.9809,
+      "step": 740
+    },
+    {
+      "epoch": 0.04922067268252666,
+      "grad_norm": 0.2672717571258545,
+      "learning_rate": 0.0002,
+      "loss": 0.8859,
+      "step": 750
+    },
+    {
+      "epoch": 0.04987694831829368,
+      "grad_norm": 0.27246803045272827,
+      "learning_rate": 0.0002,
+      "loss": 1.0251,
+      "step": 760
+    },
+    {
+      "epoch": 0.05053322395406071,
+      "grad_norm": 0.27560853958129883,
+      "learning_rate": 0.0002,
+      "loss": 0.977,
+      "step": 770
+    },
+    {
+      "epoch": 0.05118949958982773,
+      "grad_norm": 0.2449599653482437,
+      "learning_rate": 0.0002,
+      "loss": 0.9735,
+      "step": 780
+    },
+    {
+      "epoch": 0.05184577522559475,
+      "grad_norm": 0.2873939871788025,
+      "learning_rate": 0.0002,
+      "loss": 0.9765,
+      "step": 790
+    },
+    {
+      "epoch": 0.052502050861361775,
+      "grad_norm": 0.2875595688819885,
+      "learning_rate": 0.0002,
+      "loss": 0.9647,
+      "step": 800
+    },
+    {
+      "epoch": 0.053158326497128795,
+      "grad_norm": 0.4079909324645996,
+      "learning_rate": 0.0002,
+      "loss": 1.0379,
+      "step": 810
+    },
+    {
+      "epoch": 0.053814602132895815,
+      "grad_norm": 0.2733079195022583,
+      "learning_rate": 0.0002,
+      "loss": 0.9782,
+      "step": 820
+    },
+    {
+      "epoch": 0.054470877768662836,
+      "grad_norm": 0.2624184489250183,
+      "learning_rate": 0.0002,
+      "loss": 0.98,
+      "step": 830
+    },
+    {
+      "epoch": 0.05512715340442986,
+      "grad_norm": 0.279851496219635,
+      "learning_rate": 0.0002,
+      "loss": 1.0212,
+      "step": 840
+    },
+    {
+      "epoch": 0.05578342904019688,
+      "grad_norm": 0.2573111951351166,
+      "learning_rate": 0.0002,
+      "loss": 0.8994,
+      "step": 850
+    },
+    {
+      "epoch": 0.0564397046759639,
+      "grad_norm": 0.28704535961151123,
+      "learning_rate": 0.0002,
+      "loss": 0.9823,
+      "step": 860
+    },
+    {
+      "epoch": 0.05709598031173093,
+      "grad_norm": 0.2664150297641754,
+      "learning_rate": 0.0002,
+      "loss": 0.9744,
+      "step": 870
+    },
+    {
+      "epoch": 0.05775225594749795,
+      "grad_norm": 0.2858041822910309,
+      "learning_rate": 0.0002,
+      "loss": 0.9379,
+      "step": 880
+    },
+    {
+      "epoch": 0.05840853158326497,
+      "grad_norm": 0.2394150048494339,
+      "learning_rate": 0.0002,
+      "loss": 0.9453,
+      "step": 890
+    },
+    {
+      "epoch": 0.05906480721903199,
+      "grad_norm": 0.30714845657348633,
+      "learning_rate": 0.0002,
+      "loss": 0.9372,
+      "step": 900
+    },
+    {
+      "epoch": 0.05972108285479902,
+      "grad_norm": 0.2740330994129181,
+      "learning_rate": 0.0002,
+      "loss": 0.9677,
+      "step": 910
+    },
+    {
+      "epoch": 0.06037735849056604,
+      "grad_norm": 0.33505478501319885,
+      "learning_rate": 0.0002,
+      "loss": 1.0112,
+      "step": 920
+    },
+    {
+      "epoch": 0.06103363412633306,
+      "grad_norm": 0.2945438325405121,
+      "learning_rate": 0.0002,
+      "loss": 0.986,
+      "step": 930
+    },
+    {
+      "epoch": 0.06168990976210008,
+      "grad_norm": 0.24749146401882172,
+      "learning_rate": 0.0002,
+      "loss": 0.9498,
+      "step": 940
+    },
+    {
+      "epoch": 0.062346185397867106,
+      "grad_norm": 0.3315669894218445,
+      "learning_rate": 0.0002,
+      "loss": 1.0357,
+      "step": 950
+    },
+    {
+      "epoch": 0.06300246103363413,
+      "grad_norm": 0.27253520488739014,
+      "learning_rate": 0.0002,
+      "loss": 0.9698,
+      "step": 960
+    },
+    {
+      "epoch": 0.06365873666940115,
+      "grad_norm": 0.26884031295776367,
+      "learning_rate": 0.0002,
+      "loss": 0.9544,
+      "step": 970
+    },
+    {
+      "epoch": 0.06431501230516817,
+      "grad_norm": 0.286920964717865,
+      "learning_rate": 0.0002,
+      "loss": 0.9697,
+      "step": 980
+    },
+    {
+      "epoch": 0.0649712879409352,
+      "grad_norm": 0.28334400057792664,
+      "learning_rate": 0.0002,
+      "loss": 0.9704,
+      "step": 990
+    },
+    {
+      "epoch": 0.06562756357670221,
+      "grad_norm": 0.2672102749347687,
+      "learning_rate": 0.0002,
+      "loss": 0.917,
+      "step": 1000
+    },
+    {
+      "epoch": 0.06628383921246923,
+      "grad_norm": 0.3247123062610626,
+      "learning_rate": 0.0002,
+      "loss": 1.0062,
+      "step": 1010
+    },
+    {
+      "epoch": 0.06694011484823625,
+      "grad_norm": 0.259440541267395,
+      "learning_rate": 0.0002,
+      "loss": 0.9716,
+      "step": 1020
+    },
+    {
+      "epoch": 0.06759639048400329,
+      "grad_norm": 0.2795625329017639,
+      "learning_rate": 0.0002,
+      "loss": 0.9569,
+      "step": 1030
+    },
+    {
+      "epoch": 0.06825266611977031,
+      "grad_norm": 0.2784935534000397,
+      "learning_rate": 0.0002,
+      "loss": 0.9842,
+      "step": 1040
+    },
+    {
+      "epoch": 0.06890894175553733,
+      "grad_norm": 0.24605989456176758,
+      "learning_rate": 0.0002,
+      "loss": 0.9171,
+      "step": 1050
+    },
+    {
+      "epoch": 0.06956521739130435,
+      "grad_norm": 0.25421491265296936,
+      "learning_rate": 0.0002,
+      "loss": 0.9538,
+      "step": 1060
+    },
+    {
+      "epoch": 0.07022149302707137,
+      "grad_norm": 0.2693536579608917,
+      "learning_rate": 0.0002,
+      "loss": 1.0142,
+      "step": 1070
+    },
+    {
+      "epoch": 0.07087776866283839,
+      "grad_norm": 0.28166458010673523,
+      "learning_rate": 0.0002,
+      "loss": 0.9436,
+      "step": 1080
+    },
+    {
+      "epoch": 0.07153404429860541,
+      "grad_norm": 0.2752484679222107,
+      "learning_rate": 0.0002,
+      "loss": 0.9187,
+      "step": 1090
+    },
+    {
+      "epoch": 0.07219031993437244,
+      "grad_norm": 0.24141381680965424,
+      "learning_rate": 0.0002,
+      "loss": 1.0069,
+      "step": 1100
+    },
+    {
+      "epoch": 0.07284659557013946,
+      "grad_norm": 0.2347770482301712,
+      "learning_rate": 0.0002,
+      "loss": 0.9282,
+      "step": 1110
+    },
+    {
+      "epoch": 0.07350287120590648,
+      "grad_norm": 0.29999610781669617,
+      "learning_rate": 0.0002,
+      "loss": 0.978,
+      "step": 1120
+    },
+    {
+      "epoch": 0.0741591468416735,
+      "grad_norm": 0.2811068892478943,
+      "learning_rate": 0.0002,
+      "loss": 0.9763,
+      "step": 1130
+    },
+    {
+      "epoch": 0.07481542247744052,
+      "grad_norm": 0.24790801107883453,
+      "learning_rate": 0.0002,
+      "loss": 0.978,
+      "step": 1140
+    },
+    {
+      "epoch": 0.07547169811320754,
+      "grad_norm": 0.3251914978027344,
+      "learning_rate": 0.0002,
+      "loss": 1.001,
+      "step": 1150
+    },
+    {
+      "epoch": 0.07612797374897456,
+      "grad_norm": 0.2616347074508667,
+      "learning_rate": 0.0002,
+      "loss": 1.0407,
+      "step": 1160
+    },
+    {
+      "epoch": 0.07678424938474158,
+      "grad_norm": 0.2568797171115875,
+      "learning_rate": 0.0002,
+      "loss": 0.978,
+      "step": 1170
+    },
+    {
+      "epoch": 0.07744052502050862,
+      "grad_norm": 0.2693248689174652,
+      "learning_rate": 0.0002,
+      "loss": 0.9851,
+      "step": 1180
+    },
+    {
+      "epoch": 0.07809680065627564,
+      "grad_norm": 0.3270677924156189,
+      "learning_rate": 0.0002,
+      "loss": 0.996,
+      "step": 1190
+    },
+    {
+      "epoch": 0.07875307629204266,
+      "grad_norm": 0.2481861710548401,
+      "learning_rate": 0.0002,
+      "loss": 0.9751,
+      "step": 1200
+    },
+    {
+      "epoch": 0.07940935192780968,
+      "grad_norm": 0.2612398564815521,
+      "learning_rate": 0.0002,
+      "loss": 1.0047,
+      "step": 1210
+    },
+    {
+      "epoch": 0.0800656275635767,
+      "grad_norm": 0.26566916704177856,
+      "learning_rate": 0.0002,
+      "loss": 0.9789,
+      "step": 1220
+    },
+    {
+      "epoch": 0.08072190319934372,
+      "grad_norm": 0.28026407957077026,
+      "learning_rate": 0.0002,
+      "loss": 1.0468,
+      "step": 1230
+    },
+    {
+      "epoch": 0.08137817883511074,
+      "grad_norm": 0.25609225034713745,
+      "learning_rate": 0.0002,
+      "loss": 0.9651,
+      "step": 1240
+    },
+    {
+      "epoch": 0.08203445447087777,
+      "grad_norm": 0.29063138365745544,
+      "learning_rate": 0.0002,
+      "loss": 0.9648,
+      "step": 1250
+    },
+    {
+      "epoch": 0.0826907301066448,
+      "grad_norm": 0.25998231768608093,
+      "learning_rate": 0.0002,
+      "loss": 0.9302,
+      "step": 1260
+    },
+    {
+      "epoch": 0.08334700574241181,
+      "grad_norm": 0.295261412858963,
+      "learning_rate": 0.0002,
+      "loss": 0.9733,
+      "step": 1270
+    },
+    {
+      "epoch": 0.08400328137817883,
+      "grad_norm": 0.2479529082775116,
+      "learning_rate": 0.0002,
+      "loss": 0.9775,
+      "step": 1280
+    },
+    {
+      "epoch": 0.08465955701394585,
+      "grad_norm": 0.25226080417633057,
+      "learning_rate": 0.0002,
+      "loss": 0.9891,
+      "step": 1290
+    },
+    {
+      "epoch": 0.08531583264971287,
+      "grad_norm": 0.290462851524353,
+      "learning_rate": 0.0002,
+      "loss": 0.9693,
+      "step": 1300
+    },
+    {
+      "epoch": 0.0859721082854799,
+      "grad_norm": 0.2832583785057068,
+      "learning_rate": 0.0002,
+      "loss": 1.0319,
+      "step": 1310
+    },
+    {
+      "epoch": 0.08662838392124693,
+      "grad_norm": 0.2515616714954376,
+      "learning_rate": 0.0002,
+      "loss": 1.0105,
+      "step": 1320
+    },
+    {
+      "epoch": 0.08728465955701395,
+      "grad_norm": 0.28894907236099243,
+      "learning_rate": 0.0002,
+      "loss": 0.9717,
+      "step": 1330
+    },
+    {
+      "epoch": 0.08794093519278097,
+      "grad_norm": 0.2881310284137726,
+      "learning_rate": 0.0002,
+      "loss": 0.9736,
+      "step": 1340
+    },
+    {
+      "epoch": 0.08859721082854799,
+      "grad_norm": 0.2654068171977997,
+      "learning_rate": 0.0002,
+      "loss": 1.0107,
+      "step": 1350
+    },
+    {
+      "epoch": 0.08925348646431501,
+      "grad_norm": 0.2929916977882385,
+      "learning_rate": 0.0002,
+      "loss": 1.0249,
+      "step": 1360
+    },
+    {
+      "epoch": 0.08990976210008203,
+      "grad_norm": 0.27295321226119995,
+      "learning_rate": 0.0002,
+      "loss": 0.9605,
+      "step": 1370
+    },
+    {
+      "epoch": 0.09056603773584905,
+      "grad_norm": 0.27720171213150024,
+      "learning_rate": 0.0002,
+      "loss": 0.9709,
+      "step": 1380
+    },
+    {
+      "epoch": 0.09122231337161608,
+      "grad_norm": 0.2784966826438904,
+      "learning_rate": 0.0002,
+      "loss": 1.0449,
+      "step": 1390
+    },
+    {
+      "epoch": 0.0918785890073831,
+      "grad_norm": 0.25110408663749695,
+      "learning_rate": 0.0002,
+      "loss": 0.9442,
+      "step": 1400
+    },
+    {
+      "epoch": 0.09253486464315012,
+      "grad_norm": 0.2501158118247986,
+      "learning_rate": 0.0002,
+      "loss": 0.9552,
+      "step": 1410
+    },
+    {
+      "epoch": 0.09319114027891715,
+      "grad_norm": 0.3097717761993408,
+      "learning_rate": 0.0002,
+      "loss": 0.9493,
+      "step": 1420
+    },
+    {
+      "epoch": 0.09384741591468417,
+      "grad_norm": 0.257308691740036,
+      "learning_rate": 0.0002,
+      "loss": 0.9635,
+      "step": 1430
+    },
+    {
+      "epoch": 0.09450369155045119,
+      "grad_norm": 0.2885759472846985,
+      "learning_rate": 0.0002,
+      "loss": 0.9749,
+      "step": 1440
+    },
+    {
+      "epoch": 0.0951599671862182,
+      "grad_norm": 0.2938264012336731,
+      "learning_rate": 0.0002,
+      "loss": 0.9592,
+      "step": 1450
+    },
+    {
+      "epoch": 0.09581624282198524,
+      "grad_norm": 0.2543550729751587,
+      "learning_rate": 0.0002,
+      "loss": 1.0326,
+      "step": 1460
+    },
+    {
+      "epoch": 0.09647251845775226,
+      "grad_norm": 0.3020402789115906,
+      "learning_rate": 0.0002,
+      "loss": 0.9433,
+      "step": 1470
+    },
+    {
+      "epoch": 0.09712879409351928,
+      "grad_norm": 0.2818080186843872,
+      "learning_rate": 0.0002,
+      "loss": 0.9132,
+      "step": 1480
+    },
+    {
+      "epoch": 0.0977850697292863,
+      "grad_norm": 0.2420656979084015,
+      "learning_rate": 0.0002,
+      "loss": 0.9217,
+      "step": 1490
+    },
+    {
+      "epoch": 0.09844134536505332,
+      "grad_norm": 0.31427690386772156,
+      "learning_rate": 0.0002,
+      "loss": 0.938,
+      "step": 1500
+    },
+    {
+      "epoch": 0.09909762100082034,
+      "grad_norm": 0.27618250250816345,
+      "learning_rate": 0.0002,
+      "loss": 0.9234,
+      "step": 1510
+    },
+    {
+      "epoch": 0.09975389663658736,
+      "grad_norm": 0.26401254534721375,
+      "learning_rate": 0.0002,
+      "loss": 0.9019,
+      "step": 1520
+    },
+    {
+      "epoch": 0.1004101722723544,
+      "grad_norm": 0.23575739562511444,
+      "learning_rate": 0.0002,
+      "loss": 0.9992,
+      "step": 1530
+    },
+    {
+      "epoch": 0.10106644790812142,
+      "grad_norm": 0.3076087534427643,
+      "learning_rate": 0.0002,
+      "loss": 0.9315,
+      "step": 1540
+    },
+    {
+      "epoch": 0.10172272354388844,
+      "grad_norm": 0.28179317712783813,
+      "learning_rate": 0.0002,
+      "loss": 1.0138,
+      "step": 1550
+    },
+    {
+      "epoch": 0.10237899917965546,
+      "grad_norm": 0.28638190031051636,
+      "learning_rate": 0.0002,
+      "loss": 0.9587,
+      "step": 1560
+    },
+    {
+      "epoch": 0.10303527481542248,
+      "grad_norm": 0.24788478016853333,
+      "learning_rate": 0.0002,
+      "loss": 0.9576,
+      "step": 1570
+    },
+    {
+      "epoch": 0.1036915504511895,
+      "grad_norm": 0.2987830340862274,
+      "learning_rate": 0.0002,
+      "loss": 0.9523,
+      "step": 1580
+    },
+    {
+      "epoch": 0.10434782608695652,
+      "grad_norm": 0.3252484202384949,
+      "learning_rate": 0.0002,
+      "loss": 0.9782,
+      "step": 1590
+    },
+    {
+      "epoch": 0.10500410172272355,
+      "grad_norm": 0.2763068377971649,
+      "learning_rate": 0.0002,
+      "loss": 0.9378,
+      "step": 1600
+    },
+    {
+      "epoch": 0.10566037735849057,
+      "grad_norm": 0.3219335973262787,
+      "learning_rate": 0.0002,
+      "loss": 0.9491,
+      "step": 1610
+    },
+    {
+      "epoch": 0.10631665299425759,
+      "grad_norm": 0.2708939015865326,
+      "learning_rate": 0.0002,
+      "loss": 0.9181,
+      "step": 1620
+    },
+    {
+      "epoch": 0.10697292863002461,
+      "grad_norm": 0.29320818185806274,
+      "learning_rate": 0.0002,
+      "loss": 1.0008,
+      "step": 1630
+    },
+    {
+      "epoch": 0.10762920426579163,
+      "grad_norm": 0.2731851041316986,
+      "learning_rate": 0.0002,
+      "loss": 0.9593,
+      "step": 1640
+    },
+    {
+      "epoch": 0.10828547990155865,
+      "grad_norm": 0.29563331604003906,
+      "learning_rate": 0.0002,
+      "loss": 0.9915,
+      "step": 1650
+    },
+    {
+      "epoch": 0.10894175553732567,
+      "grad_norm": 0.33387669920921326,
+      "learning_rate": 0.0002,
+      "loss": 0.9543,
+      "step": 1660
+    },
+    {
+      "epoch": 0.1095980311730927,
+      "grad_norm": 0.3157867193222046,
+      "learning_rate": 0.0002,
+      "loss": 0.941,
+      "step": 1670
+    },
+    {
+      "epoch": 0.11025430680885973,
+      "grad_norm": 0.24688409268856049,
+      "learning_rate": 0.0002,
+      "loss": 0.9107,
+      "step": 1680
+    },
+    {
+      "epoch": 0.11091058244462675,
+      "grad_norm": 0.2868671417236328,
+      "learning_rate": 0.0002,
+      "loss": 0.9392,
+      "step": 1690
+    },
+    {
+      "epoch": 0.11156685808039377,
+      "grad_norm": 0.2555101811885834,
+      "learning_rate": 0.0002,
+      "loss": 0.9476,
+      "step": 1700
+    },
+    {
+      "epoch": 0.11222313371616079,
+      "grad_norm": 0.3224661946296692,
+      "learning_rate": 0.0002,
+      "loss": 0.9261,
+      "step": 1710
+    },
+    {
+      "epoch": 0.1128794093519278,
+      "grad_norm": 0.28927505016326904,
+      "learning_rate": 0.0002,
+      "loss": 0.9372,
+      "step": 1720
+    },
+    {
+      "epoch": 0.11353568498769483,
+      "grad_norm": 0.25575608015060425,
+      "learning_rate": 0.0002,
+      "loss": 0.9237,
+      "step": 1730
+    },
+    {
+      "epoch": 0.11419196062346186,
+      "grad_norm": 0.27490100264549255,
+      "learning_rate": 0.0002,
+      "loss": 0.9526,
+      "step": 1740
+    },
+    {
+      "epoch": 0.11484823625922888,
+      "grad_norm": 0.28235989809036255,
+      "learning_rate": 0.0002,
+      "loss": 1.0311,
+      "step": 1750
+    },
+    {
+      "epoch": 0.1155045118949959,
+      "grad_norm": 0.2636060416698456,
+      "learning_rate": 0.0002,
+      "loss": 0.9112,
+      "step": 1760
+    },
+    {
+      "epoch": 0.11616078753076292,
+      "grad_norm": 0.2642059624195099,
+      "learning_rate": 0.0002,
+      "loss": 0.9249,
+      "step": 1770
+    },
+    {
+      "epoch": 0.11681706316652994,
+      "grad_norm": 0.261807918548584,
+      "learning_rate": 0.0002,
+      "loss": 0.964,
+      "step": 1780
+    },
+    {
+      "epoch": 0.11747333880229696,
+      "grad_norm": 0.24618281424045563,
+      "learning_rate": 0.0002,
+      "loss": 0.9473,
+      "step": 1790
+    },
+    {
+      "epoch": 0.11812961443806398,
+      "grad_norm": 0.2700425982475281,
+      "learning_rate": 0.0002,
+      "loss": 0.9452,
+      "step": 1800
+    },
+    {
+      "epoch": 0.11878589007383102,
+      "grad_norm": 0.22687454521656036,
+      "learning_rate": 0.0002,
+      "loss": 0.9687,
+      "step": 1810
+    },
+    {
+      "epoch": 0.11944216570959804,
+      "grad_norm": 0.2843499481678009,
+      "learning_rate": 0.0002,
+      "loss": 0.9099,
+      "step": 1820
+    },
+    {
+      "epoch": 0.12009844134536506,
+      "grad_norm": 0.23511677980422974,
+      "learning_rate": 0.0002,
+      "loss": 0.9295,
+      "step": 1830
+    },
+    {
+      "epoch": 0.12075471698113208,
+      "grad_norm": 0.34054702520370483,
+      "learning_rate": 0.0002,
+      "loss": 0.998,
+      "step": 1840
+    },
+    {
+      "epoch": 0.1214109926168991,
+      "grad_norm": 0.32845374941825867,
+      "learning_rate": 0.0002,
+      "loss": 0.9343,
+      "step": 1850
+    },
+    {
+      "epoch": 0.12206726825266612,
+      "grad_norm": 0.2843034267425537,
+      "learning_rate": 0.0002,
+      "loss": 0.983,
+      "step": 1860
+    },
+    {
+      "epoch": 0.12272354388843314,
+      "grad_norm": 0.2824552357196808,
+      "learning_rate": 0.0002,
+      "loss": 0.963,
+      "step": 1870
+    },
+    {
+      "epoch": 0.12337981952420016,
+      "grad_norm": 0.26878267526626587,
+      "learning_rate": 0.0002,
+      "loss": 0.9875,
+      "step": 1880
+    },
+    {
+      "epoch": 0.12403609515996719,
+      "grad_norm": 0.31207871437072754,
+      "learning_rate": 0.0002,
+      "loss": 0.9784,
+      "step": 1890
+    },
+    {
+      "epoch": 0.12469237079573421,
+      "grad_norm": 0.2617819905281067,
+      "learning_rate": 0.0002,
+      "loss": 0.9497,
+      "step": 1900
+    },
+    {
+      "epoch": 0.12534864643150123,
+      "grad_norm": 0.2652885615825653,
+      "learning_rate": 0.0002,
+      "loss": 0.8948,
+      "step": 1910
+    },
+    {
+      "epoch": 0.12600492206726827,
+      "grad_norm": 0.29473352432250977,
+      "learning_rate": 0.0002,
+      "loss": 0.9534,
+      "step": 1920
+    },
+    {
+      "epoch": 0.12666119770303527,
+      "grad_norm": 0.3101664185523987,
+      "learning_rate": 0.0002,
+      "loss": 0.9695,
+      "step": 1930
+    },
+    {
+      "epoch": 0.1273174733388023,
+      "grad_norm": 0.2714068293571472,
+      "learning_rate": 0.0002,
+      "loss": 0.9928,
+      "step": 1940
+    },
+    {
+      "epoch": 0.1279737489745693,
+      "grad_norm": 0.2684655487537384,
+      "learning_rate": 0.0002,
+      "loss": 1.0293,
+      "step": 1950
+    },
+    {
+      "epoch": 0.12863002461033635,
+      "grad_norm": 0.2543509900569916,
+      "learning_rate": 0.0002,
+      "loss": 0.9236,
+      "step": 1960
+    },
+    {
+      "epoch": 0.12928630024610335,
+      "grad_norm": 0.25268790125846863,
+      "learning_rate": 0.0002,
+      "loss": 0.9598,
+      "step": 1970
+    },
+    {
+      "epoch": 0.1299425758818704,
+      "grad_norm": 0.27713078260421753,
+      "learning_rate": 0.0002,
+      "loss": 0.967,
+      "step": 1980
+    },
+    {
+      "epoch": 0.13059885151763742,
+      "grad_norm": 0.3185126483440399,
+      "learning_rate": 0.0002,
+      "loss": 0.9411,
+      "step": 1990
+    },
+    {
+      "epoch": 0.13125512715340443,
+      "grad_norm": 0.26718857884407043,
+      "learning_rate": 0.0002,
+      "loss": 0.9864,
+      "step": 2000
+    },
+    {
+      "epoch": 0.13191140278917146,
+      "grad_norm": 0.2843841314315796,
+      "learning_rate": 0.0002,
+      "loss": 0.9728,
+      "step": 2010
+    },
+    {
+      "epoch": 0.13256767842493847,
+      "grad_norm": 0.3013055622577667,
+      "learning_rate": 0.0002,
+      "loss": 0.993,
+      "step": 2020
+    },
+    {
+      "epoch": 0.1332239540607055,
+      "grad_norm": 0.27985867857933044,
+      "learning_rate": 0.0002,
+      "loss": 0.9437,
+      "step": 2030
+    },
+    {
+      "epoch": 0.1338802296964725,
+      "grad_norm": 0.2501908242702484,
+      "learning_rate": 0.0002,
+      "loss": 0.905,
+      "step": 2040
+    },
+    {
+      "epoch": 0.13453650533223954,
+      "grad_norm": 0.441340833902359,
+      "learning_rate": 0.0002,
+      "loss": 0.9603,
+      "step": 2050
+    },
+    {
+      "epoch": 0.13519278096800658,
+      "grad_norm": 2.568060874938965,
+      "learning_rate": 0.0002,
+      "loss": 0.9166,
+      "step": 2060
+    },
+    {
+      "epoch": 0.13584905660377358,
+      "grad_norm": 0.2505454421043396,
+      "learning_rate": 0.0002,
+      "loss": 0.9803,
+      "step": 2070
+    },
+    {
+      "epoch": 0.13650533223954062,
+      "grad_norm": 0.33997446298599243,
+      "learning_rate": 0.0002,
+      "loss": 1.0159,
+      "step": 2080
+    },
+    {
+      "epoch": 0.13716160787530762,
+      "grad_norm": 0.28161343932151794,
+      "learning_rate": 0.0002,
+      "loss": 0.908,
+      "step": 2090
+    },
+    {
+      "epoch": 0.13781788351107466,
+      "grad_norm": 0.28142687678337097,
+      "learning_rate": 0.0002,
+      "loss": 0.9709,
+      "step": 2100
+    },
+    {
+      "epoch": 0.13847415914684166,
+      "grad_norm": 0.3137170076370239,
+      "learning_rate": 0.0002,
+      "loss": 0.9716,
+      "step": 2110
+    },
+    {
+      "epoch": 0.1391304347826087,
+      "grad_norm": 0.3072240352630615,
+      "learning_rate": 0.0002,
+      "loss": 0.9698,
+      "step": 2120
+    },
+    {
+      "epoch": 0.13978671041837573,
+      "grad_norm": 0.24328380823135376,
+      "learning_rate": 0.0002,
+      "loss": 0.9979,
+      "step": 2130
+    },
+    {
+      "epoch": 0.14044298605414274,
+      "grad_norm": 0.3065047860145569,
+      "learning_rate": 0.0002,
+      "loss": 0.9549,
+      "step": 2140
+    },
+    {
+      "epoch": 0.14109926168990977,
+      "grad_norm": 0.34212175011634827,
+      "learning_rate": 0.0002,
+      "loss": 0.9897,
+      "step": 2150
+    },
+    {
+      "epoch": 0.14175553732567678,
+      "grad_norm": 0.27491796016693115,
+      "learning_rate": 0.0002,
+      "loss": 0.9601,
+      "step": 2160
+    },
+    {
+      "epoch": 0.1424118129614438,
+      "grad_norm": 0.26518693566322327,
+      "learning_rate": 0.0002,
+      "loss": 0.9928,
+      "step": 2170
+    },
+    {
+      "epoch": 0.14306808859721082,
+      "grad_norm": 0.28350934386253357,
+      "learning_rate": 0.0002,
+      "loss": 0.9385,
+      "step": 2180
+    },
+    {
+      "epoch": 0.14372436423297785,
+      "grad_norm": 0.3287768065929413,
+      "learning_rate": 0.0002,
+      "loss": 0.9878,
+      "step": 2190
+    },
+    {
+      "epoch": 0.1443806398687449,
+      "grad_norm": 0.26362666487693787,
+      "learning_rate": 0.0002,
+      "loss": 0.9192,
+      "step": 2200
+    },
+    {
+      "epoch": 0.1450369155045119,
+      "grad_norm": 0.31169235706329346,
+      "learning_rate": 0.0002,
+      "loss": 0.9804,
+      "step": 2210
+    },
+    {
+      "epoch": 0.14569319114027893,
+      "grad_norm": 0.258667528629303,
+      "learning_rate": 0.0002,
+      "loss": 0.9857,
+      "step": 2220
+    },
+    {
+      "epoch": 0.14634946677604593,
+      "grad_norm": 0.32289111614227295,
+      "learning_rate": 0.0002,
+      "loss": 0.9898,
+      "step": 2230
+    },
+    {
+      "epoch": 0.14700574241181297,
+      "grad_norm": 0.3344270586967468,
+      "learning_rate": 0.0002,
+      "loss": 1.0051,
+      "step": 2240
+    },
+    {
+      "epoch": 0.14766201804757997,
+      "grad_norm": 0.3001033067703247,
+      "learning_rate": 0.0002,
+      "loss": 0.9337,
+      "step": 2250
+    },
+    {
+      "epoch": 0.148318293683347,
+      "grad_norm": 0.2718261778354645,
+      "learning_rate": 0.0002,
+      "loss": 0.9788,
+      "step": 2260
+    },
+    {
+      "epoch": 0.14897456931911401,
+      "grad_norm": 0.3059164583683014,
+      "learning_rate": 0.0002,
+      "loss": 0.9801,
+      "step": 2270
+    },
+    {
+      "epoch": 0.14963084495488105,
+      "grad_norm": 0.2939850389957428,
+      "learning_rate": 0.0002,
+      "loss": 0.8914,
+      "step": 2280
+    },
+    {
+      "epoch": 0.15028712059064808,
+      "grad_norm": 0.2803564965724945,
+      "learning_rate": 0.0002,
+      "loss": 0.951,
+      "step": 2290
+    },
+    {
+      "epoch": 0.1509433962264151,
+      "grad_norm": 0.2843068242073059,
+      "learning_rate": 0.0002,
+      "loss": 0.9131,
+      "step": 2300
+    },
+    {
+      "epoch": 0.15159967186218212,
+      "grad_norm": 0.28517085313796997,
+      "learning_rate": 0.0002,
+      "loss": 0.9721,
+      "step": 2310
+    },
+    {
+      "epoch": 0.15225594749794913,
+      "grad_norm": 0.291030615568161,
+      "learning_rate": 0.0002,
+      "loss": 0.9402,
+      "step": 2320
+    },
+    {
+      "epoch": 0.15291222313371616,
+      "grad_norm": 0.2712008059024811,
+      "learning_rate": 0.0002,
+      "loss": 0.9509,
+      "step": 2330
+    },
+    {
+      "epoch": 0.15356849876948317,
+      "grad_norm": 0.3357657790184021,
+      "learning_rate": 0.0002,
+      "loss": 0.9406,
+      "step": 2340
+    },
+    {
+      "epoch": 0.1542247744052502,
+      "grad_norm": 0.28412291407585144,
+      "learning_rate": 0.0002,
+      "loss": 0.961,
+      "step": 2350
+    },
+    {
+      "epoch": 0.15488105004101724,
+      "grad_norm": 0.3217862844467163,
+      "learning_rate": 0.0002,
+      "loss": 0.9254,
+      "step": 2360
+    },
+    {
+      "epoch": 0.15553732567678424,
+      "grad_norm": 0.32076528668403625,
+      "learning_rate": 0.0002,
+      "loss": 0.9172,
+      "step": 2370
+    },
+    {
+      "epoch": 0.15619360131255128,
+      "grad_norm": 0.3490257263183594,
+      "learning_rate": 0.0002,
+      "loss": 0.9459,
+      "step": 2380
+    },
+    {
+      "epoch": 0.15684987694831828,
+      "grad_norm": 0.28253331780433655,
+      "learning_rate": 0.0002,
+      "loss": 0.9051,
+      "step": 2390
+    },
+    {
+      "epoch": 0.15750615258408532,
+      "grad_norm": 0.2875654399394989,
+      "learning_rate": 0.0002,
+      "loss": 0.9597,
+      "step": 2400
+    },
+    {
+      "epoch": 0.15816242821985232,
+      "grad_norm": 0.3347921669483185,
+      "learning_rate": 0.0002,
+      "loss": 0.9164,
+      "step": 2410
+    },
+    {
+      "epoch": 0.15881870385561936,
+      "grad_norm": 0.2624322474002838,
+      "learning_rate": 0.0002,
+      "loss": 1.0023,
+      "step": 2420
+    },
+    {
+      "epoch": 0.1594749794913864,
+      "grad_norm": 0.29683780670166016,
+      "learning_rate": 0.0002,
+      "loss": 0.9751,
+      "step": 2430
+    },
+    {
+      "epoch": 0.1601312551271534,
+      "grad_norm": 0.29237303137779236,
+      "learning_rate": 0.0002,
+      "loss": 0.967,
+      "step": 2440
+    },
+    {
+      "epoch": 0.16078753076292043,
+      "grad_norm": 0.3541257977485657,
+      "learning_rate": 0.0002,
+      "loss": 0.9754,
+      "step": 2450
+    },
+    {
+      "epoch": 0.16144380639868744,
+      "grad_norm": 0.28203415870666504,
+      "learning_rate": 0.0002,
+      "loss": 0.8977,
+      "step": 2460
+    },
+    {
+      "epoch": 0.16210008203445447,
+      "grad_norm": 0.322329044342041,
+      "learning_rate": 0.0002,
+      "loss": 0.9489,
+      "step": 2470
+    },
+    {
+      "epoch": 0.16275635767022148,
+      "grad_norm": 0.27731558680534363,
+      "learning_rate": 0.0002,
+      "loss": 0.9224,
+      "step": 2480
+    },
+    {
+      "epoch": 0.16341263330598851,
+      "grad_norm": 0.30348825454711914,
+      "learning_rate": 0.0002,
+      "loss": 0.9773,
+      "step": 2490
+    },
+    {
+      "epoch": 0.16406890894175555,
+      "grad_norm": 0.29504773020744324,
+      "learning_rate": 0.0002,
+      "loss": 0.9318,
+      "step": 2500
+    },
+    {
+      "epoch": 0.16472518457752255,
+      "grad_norm": 0.2977028489112854,
+      "learning_rate": 0.0002,
+      "loss": 0.8931,
+      "step": 2510
+    },
+    {
+      "epoch": 0.1653814602132896,
+      "grad_norm": 0.3172457218170166,
+      "learning_rate": 0.0002,
+      "loss": 0.953,
+      "step": 2520
+    },
+    {
+      "epoch": 0.1660377358490566,
+      "grad_norm": 0.3320612907409668,
+      "learning_rate": 0.0002,
+      "loss": 0.9786,
+      "step": 2530
+    },
+    {
+      "epoch": 0.16669401148482363,
+      "grad_norm": 0.2473023235797882,
+      "learning_rate": 0.0002,
+      "loss": 0.895,
+      "step": 2540
+    },
+    {
+      "epoch": 0.16735028712059064,
+      "grad_norm": 0.3250006139278412,
+      "learning_rate": 0.0002,
+      "loss": 0.9692,
+      "step": 2550
+    },
+    {
+      "epoch": 0.16800656275635767,
+      "grad_norm": 0.2940689027309418,
+      "learning_rate": 0.0002,
+      "loss": 0.956,
+      "step": 2560
+    },
+    {
+      "epoch": 0.1686628383921247,
+      "grad_norm": 0.34614887833595276,
+      "learning_rate": 0.0002,
+      "loss": 0.8926,
+      "step": 2570
+    },
+    {
+      "epoch": 0.1693191140278917,
+      "grad_norm": 0.3513164520263672,
+      "learning_rate": 0.0002,
+      "loss": 1.0153,
+      "step": 2580
+    },
+    {
+      "epoch": 0.16997538966365874,
+      "grad_norm": 0.2530531585216522,
+      "learning_rate": 0.0002,
+      "loss": 0.9608,
+      "step": 2590
+    },
+    {
+      "epoch": 0.17063166529942575,
+      "grad_norm": 0.3028896152973175,
+      "learning_rate": 0.0002,
+      "loss": 0.9876,
+      "step": 2600
+    },
+    {
+      "epoch": 0.17128794093519278,
+      "grad_norm": 0.304739773273468,
+      "learning_rate": 0.0002,
+      "loss": 0.9415,
+      "step": 2610
+    },
+    {
+      "epoch": 0.1719442165709598,
+      "grad_norm": 0.3360660672187805,
+      "learning_rate": 0.0002,
+      "loss": 0.9857,
+      "step": 2620
+    },
+    {
+      "epoch": 0.17260049220672682,
+      "grad_norm": 0.24901753664016724,
+      "learning_rate": 0.0002,
+      "loss": 0.9173,
+      "step": 2630
+    },
+    {
+      "epoch": 0.17325676784249386,
+      "grad_norm": 0.32212400436401367,
+      "learning_rate": 0.0002,
+      "loss": 0.9351,
+      "step": 2640
+    },
+    {
+      "epoch": 0.17391304347826086,
+      "grad_norm": 0.2780437767505646,
+      "learning_rate": 0.0002,
+      "loss": 0.9284,
+      "step": 2650
+    },
+    {
+      "epoch": 0.1745693191140279,
+      "grad_norm": 0.25902262330055237,
+      "learning_rate": 0.0002,
+      "loss": 0.9368,
+      "step": 2660
+    },
+    {
+      "epoch": 0.1752255947497949,
+      "grad_norm": 0.29171422123908997,
+      "learning_rate": 0.0002,
+      "loss": 0.9378,
+      "step": 2670
+    },
+    {
+      "epoch": 0.17588187038556194,
+      "grad_norm": 0.2586783766746521,
+      "learning_rate": 0.0002,
+      "loss": 0.9525,
+      "step": 2680
+    },
+    {
+      "epoch": 0.17653814602132895,
+      "grad_norm": 0.25609949231147766,
+      "learning_rate": 0.0002,
+      "loss": 0.9367,
+      "step": 2690
+    },
+    {
+      "epoch": 0.17719442165709598,
+      "grad_norm": 0.29789483547210693,
+      "learning_rate": 0.0002,
+      "loss": 0.8675,
+      "step": 2700
+    },
+    {
+      "epoch": 0.177850697292863,
+      "grad_norm": 0.2938411235809326,
+      "learning_rate": 0.0002,
+      "loss": 0.9918,
+      "step": 2710
+    },
+    {
+      "epoch": 0.17850697292863002,
+      "grad_norm": 0.31436532735824585,
+      "learning_rate": 0.0002,
+      "loss": 0.9824,
+      "step": 2720
+    },
+    {
+      "epoch": 0.17916324856439705,
+      "grad_norm": 0.31310203671455383,
+      "learning_rate": 0.0002,
+      "loss": 1.0094,
+      "step": 2730
+    },
+    {
+      "epoch": 0.17981952420016406,
+      "grad_norm": 0.2695367634296417,
+      "learning_rate": 0.0002,
+      "loss": 0.9105,
+      "step": 2740
+    },
+    {
+      "epoch": 0.1804757998359311,
+      "grad_norm": 0.32590436935424805,
+      "learning_rate": 0.0002,
+      "loss": 0.9216,
+      "step": 2750
+    },
+    {
+      "epoch": 0.1811320754716981,
+      "grad_norm": 0.32683756947517395,
+      "learning_rate": 0.0002,
+      "loss": 0.923,
+      "step": 2760
+    },
+    {
+      "epoch": 0.18178835110746513,
+      "grad_norm": 0.33043643832206726,
+      "learning_rate": 0.0002,
+      "loss": 0.938,
+      "step": 2770
+    },
+    {
+      "epoch": 0.18244462674323217,
+      "grad_norm": 0.3137816786766052,
+      "learning_rate": 0.0002,
+      "loss": 0.937,
+      "step": 2780
+    },
+    {
+      "epoch": 0.18310090237899918,
+      "grad_norm": 0.33037737011909485,
+      "learning_rate": 0.0002,
+      "loss": 0.8923,
+      "step": 2790
+    },
+    {
+      "epoch": 0.1837571780147662,
+      "grad_norm": 0.27679184079170227,
+      "learning_rate": 0.0002,
+      "loss": 0.9179,
+      "step": 2800
+    },
+    {
+      "epoch": 0.18441345365053322,
+      "grad_norm": 0.42975902557373047,
+      "learning_rate": 0.0002,
+      "loss": 0.9047,
+      "step": 2810
+    },
+    {
+      "epoch": 0.18506972928630025,
+      "grad_norm": 0.2915041744709015,
+      "learning_rate": 0.0002,
+      "loss": 0.9431,
+      "step": 2820
+    },
+    {
+      "epoch": 0.18572600492206726,
+      "grad_norm": 0.315197229385376,
+      "learning_rate": 0.0002,
+      "loss": 0.9573,
+      "step": 2830
+    },
+    {
+      "epoch": 0.1863822805578343,
+      "grad_norm": 0.31129002571105957,
+      "learning_rate": 0.0002,
+      "loss": 0.9909,
+      "step": 2840
+    },
+    {
+      "epoch": 0.18703855619360132,
+      "grad_norm": 0.2783232033252716,
+      "learning_rate": 0.0002,
+      "loss": 0.9222,
+      "step": 2850
+    },
+    {
+      "epoch": 0.18769483182936833,
+      "grad_norm": 0.34501492977142334,
+      "learning_rate": 0.0002,
+      "loss": 0.9347,
+      "step": 2860
+    },
+    {
+      "epoch": 0.18835110746513536,
+      "grad_norm": 0.32616767287254333,
+      "learning_rate": 0.0002,
+      "loss": 0.949,
+      "step": 2870
+    },
+    {
+      "epoch": 0.18900738310090237,
+      "grad_norm": 0.25213682651519775,
+      "learning_rate": 0.0002,
+      "loss": 0.9909,
+      "step": 2880
+    },
+    {
+      "epoch": 0.1896636587366694,
+      "grad_norm": 0.2745118737220764,
+      "learning_rate": 0.0002,
+      "loss": 0.892,
+      "step": 2890
+    },
+    {
+      "epoch": 0.1903199343724364,
+      "grad_norm": 0.26175656914711,
+      "learning_rate": 0.0002,
+      "loss": 0.9737,
+      "step": 2900
+    },
+    {
+      "epoch": 0.19097621000820345,
+      "grad_norm": 0.3019673228263855,
+      "learning_rate": 0.0002,
+      "loss": 0.8853,
+      "step": 2910
+    },
+    {
+      "epoch": 0.19163248564397048,
+      "grad_norm": 0.30918899178504944,
+      "learning_rate": 0.0002,
+      "loss": 1.0016,
+      "step": 2920
+    },
+    {
+      "epoch": 0.19228876127973749,
+      "grad_norm": 0.320893794298172,
+      "learning_rate": 0.0002,
+      "loss": 0.9301,
+      "step": 2930
+    },
+    {
+      "epoch": 0.19294503691550452,
+      "grad_norm": 0.2736368477344513,
+      "learning_rate": 0.0002,
+      "loss": 0.9592,
+      "step": 2940
+    },
+    {
+      "epoch": 0.19360131255127153,
+      "grad_norm": 0.28817251324653625,
+      "learning_rate": 0.0002,
+      "loss": 0.9334,
+      "step": 2950
+    },
+    {
+      "epoch": 0.19425758818703856,
+      "grad_norm": 0.26863837242126465,
+      "learning_rate": 0.0002,
+      "loss": 0.9484,
+      "step": 2960
+    },
+    {
+      "epoch": 0.19491386382280557,
+      "grad_norm": 0.28012585639953613,
+      "learning_rate": 0.0002,
+      "loss": 0.9466,
+      "step": 2970
+    },
+    {
+      "epoch": 0.1955701394585726,
+      "grad_norm": 0.2929932773113251,
+      "learning_rate": 0.0002,
+      "loss": 0.9623,
+      "step": 2980
+    },
+    {
+      "epoch": 0.19622641509433963,
+      "grad_norm": 0.29878300428390503,
+      "learning_rate": 0.0002,
+      "loss": 0.9406,
+      "step": 2990
+    },
+    {
+      "epoch": 0.19688269073010664,
+      "grad_norm": 0.2573733329772949,
+      "learning_rate": 0.0002,
+      "loss": 0.9714,
+      "step": 3000
+    },
+    {
+      "epoch": 0.19753896636587367,
+      "grad_norm": 0.31736820936203003,
+      "learning_rate": 0.0002,
+      "loss": 0.9763,
+      "step": 3010
+    },
+    {
+      "epoch": 0.19819524200164068,
+      "grad_norm": 0.2864682972431183,
+      "learning_rate": 0.0002,
+      "loss": 1.0111,
+      "step": 3020
+    },
+    {
+      "epoch": 0.19885151763740772,
+      "grad_norm": 0.3161790668964386,
+      "learning_rate": 0.0002,
+      "loss": 0.9193,
+      "step": 3030
+    },
+    {
+      "epoch": 0.19950779327317472,
+      "grad_norm": 0.37538009881973267,
+      "learning_rate": 0.0002,
+      "loss": 0.9298,
+      "step": 3040
+    },
+    {
+      "epoch": 0.20016406890894176,
+      "grad_norm": 0.2850522994995117,
+      "learning_rate": 0.0002,
+      "loss": 0.9329,
+      "step": 3050
+    },
+    {
+      "epoch": 0.2008203445447088,
+      "grad_norm": 0.36680465936660767,
+      "learning_rate": 0.0002,
+      "loss": 1.0777,
+      "step": 3060
+    },
+    {
+      "epoch": 0.2014766201804758,
+      "grad_norm": 0.2879831790924072,
+      "learning_rate": 0.0002,
+      "loss": 0.9042,
+      "step": 3070
+    },
+    {
+      "epoch": 0.20213289581624283,
+      "grad_norm": 0.29652512073516846,
+      "learning_rate": 0.0002,
+      "loss": 0.8925,
+      "step": 3080
+    },
+    {
+      "epoch": 0.20278917145200984,
+      "grad_norm": 0.3131853938102722,
+      "learning_rate": 0.0002,
+      "loss": 0.9928,
+      "step": 3090
+    },
+    {
+      "epoch": 0.20344544708777687,
+      "grad_norm": 0.34605351090431213,
+      "learning_rate": 0.0002,
+      "loss": 0.9968,
+      "step": 3100
+    },
+    {
+      "epoch": 0.20410172272354388,
+      "grad_norm": 0.30542251467704773,
+      "learning_rate": 0.0002,
+      "loss": 0.9701,
+      "step": 3110
+    },
+    {
+      "epoch": 0.2047579983593109,
+      "grad_norm": 0.28131186962127686,
+      "learning_rate": 0.0002,
+      "loss": 0.9483,
+      "step": 3120
+    },
+    {
+      "epoch": 0.20541427399507795,
+      "grad_norm": 0.257859468460083,
+      "learning_rate": 0.0002,
+      "loss": 0.9332,
+      "step": 3130
+    },
+    {
+      "epoch": 0.20607054963084495,
+      "grad_norm": 0.32994887232780457,
+      "learning_rate": 0.0002,
+      "loss": 0.9771,
+      "step": 3140
+    },
+    {
+      "epoch": 0.20672682526661199,
+      "grad_norm": 0.26293760538101196,
+      "learning_rate": 0.0002,
+      "loss": 0.9368,
+      "step": 3150
+    },
+    {
+      "epoch": 0.207383100902379,
+      "grad_norm": 0.29514846205711365,
+      "learning_rate": 0.0002,
+      "loss": 0.9801,
+      "step": 3160
+    },
+    {
+      "epoch": 0.20803937653814603,
+      "grad_norm": 0.36102691292762756,
+      "learning_rate": 0.0002,
+      "loss": 0.9879,
+      "step": 3170
+    },
+    {
+      "epoch": 0.20869565217391303,
+      "grad_norm": 0.3298998475074768,
+      "learning_rate": 0.0002,
+      "loss": 0.9322,
+      "step": 3180
+    },
+    {
+      "epoch": 0.20935192780968007,
+      "grad_norm": 0.29835769534111023,
+      "learning_rate": 0.0002,
+      "loss": 0.9456,
+      "step": 3190
+    },
+    {
+      "epoch": 0.2100082034454471,
+      "grad_norm": 0.3438013792037964,
+      "learning_rate": 0.0002,
+      "loss": 1.0199,
+      "step": 3200
+    },
+    {
+      "epoch": 0.2106644790812141,
+      "grad_norm": 0.2945845127105713,
+      "learning_rate": 0.0002,
+      "loss": 0.9104,
+      "step": 3210
+    },
+    {
+      "epoch": 0.21132075471698114,
+      "grad_norm": 0.3173643946647644,
+      "learning_rate": 0.0002,
+      "loss": 0.9231,
+      "step": 3220
+    },
+    {
+      "epoch": 0.21197703035274815,
+      "grad_norm": 0.30580341815948486,
+      "learning_rate": 0.0002,
+      "loss": 0.892,
+      "step": 3230
+    },
+    {
+      "epoch": 0.21263330598851518,
+      "grad_norm": 0.32913172245025635,
+      "learning_rate": 0.0002,
+      "loss": 0.9458,
+      "step": 3240
+    },
+    {
+      "epoch": 0.2132895816242822,
+      "grad_norm": 0.2739659249782562,
+      "learning_rate": 0.0002,
+      "loss": 0.9475,
+      "step": 3250
+    },
+    {
+      "epoch": 0.21394585726004922,
+      "grad_norm": 0.3016273081302643,
+      "learning_rate": 0.0002,
+      "loss": 0.9185,
+      "step": 3260
+    },
+    {
+      "epoch": 0.21460213289581626,
+      "grad_norm": 0.2818678319454193,
+      "learning_rate": 0.0002,
+      "loss": 0.9819,
+      "step": 3270
+    },
+    {
+      "epoch": 0.21525840853158326,
+      "grad_norm": 0.3243506848812103,
+      "learning_rate": 0.0002,
+      "loss": 0.9535,
+      "step": 3280
+    },
+    {
+      "epoch": 0.2159146841673503,
+      "grad_norm": 0.38120919466018677,
+      "learning_rate": 0.0002,
+      "loss": 0.9355,
+      "step": 3290
+    },
+    {
+      "epoch": 0.2165709598031173,
+      "grad_norm": 0.3041105568408966,
+      "learning_rate": 0.0002,
+      "loss": 0.9556,
+      "step": 3300
+    },
+    {
+      "epoch": 0.21722723543888434,
+      "grad_norm": 0.2648089528083801,
+      "learning_rate": 0.0002,
+      "loss": 0.9172,
+      "step": 3310
+    },
+    {
+      "epoch": 0.21788351107465134,
+      "grad_norm": 0.324095219373703,
+      "learning_rate": 0.0002,
+      "loss": 0.9567,
+      "step": 3320
+    },
+    {
+      "epoch": 0.21853978671041838,
+      "grad_norm": 0.2796897888183594,
+      "learning_rate": 0.0002,
+      "loss": 0.9586,
+      "step": 3330
+    },
+    {
+      "epoch": 0.2191960623461854,
+      "grad_norm": 0.30163177847862244,
+      "learning_rate": 0.0002,
+      "loss": 0.9505,
+      "step": 3340
+    },
+    {
+      "epoch": 0.21985233798195242,
+      "grad_norm": 0.29213520884513855,
+      "learning_rate": 0.0002,
+      "loss": 0.9515,
+      "step": 3350
+    },
+    {
+      "epoch": 0.22050861361771945,
+      "grad_norm": 0.28203412890434265,
+      "learning_rate": 0.0002,
+      "loss": 0.9692,
+      "step": 3360
+    },
+    {
+      "epoch": 0.22116488925348646,
+      "grad_norm": 0.31636562943458557,
+      "learning_rate": 0.0002,
+      "loss": 0.94,
+      "step": 3370
+    },
+    {
+      "epoch": 0.2218211648892535,
+      "grad_norm": 0.3153952956199646,
+      "learning_rate": 0.0002,
+      "loss": 0.901,
+      "step": 3380
+    },
+    {
+      "epoch": 0.2224774405250205,
+      "grad_norm": 0.2793780267238617,
+      "learning_rate": 0.0002,
+      "loss": 0.8994,
+      "step": 3390
+    },
+    {
+      "epoch": 0.22313371616078753,
+      "grad_norm": 0.36783504486083984,
+      "learning_rate": 0.0002,
+      "loss": 0.9424,
+      "step": 3400
+    },
+    {
+      "epoch": 0.22378999179655457,
+      "grad_norm": 0.31803956627845764,
+      "learning_rate": 0.0002,
+      "loss": 0.9837,
+      "step": 3410
+    },
+    {
+      "epoch": 0.22444626743232157,
+      "grad_norm": 0.31517738103866577,
+      "learning_rate": 0.0002,
+      "loss": 0.9234,
+      "step": 3420
+    },
+    {
+      "epoch": 0.2251025430680886,
+      "grad_norm": 0.4124458432197571,
+      "learning_rate": 0.0002,
+      "loss": 0.9336,
+      "step": 3430
+    },
+    {
+      "epoch": 0.2257588187038556,
+      "grad_norm": 0.29330259561538696,
+      "learning_rate": 0.0002,
+      "loss": 0.8587,
+      "step": 3440
+    },
+    {
+      "epoch": 0.22641509433962265,
+      "grad_norm": 0.281255304813385,
+      "learning_rate": 0.0002,
+      "loss": 0.925,
+      "step": 3450
+    },
+    {
+      "epoch": 0.22707136997538965,
+      "grad_norm": 0.3765242397785187,
+      "learning_rate": 0.0002,
+      "loss": 0.9797,
+      "step": 3460
+    },
+    {
+      "epoch": 0.2277276456111567,
+      "grad_norm": 0.33658838272094727,
+      "learning_rate": 0.0002,
+      "loss": 1.0201,
+      "step": 3470
+    },
+    {
+      "epoch": 0.22838392124692372,
+      "grad_norm": 0.2809208035469055,
+      "learning_rate": 0.0002,
+      "loss": 0.9255,
+      "step": 3480
+    },
+    {
+      "epoch": 0.22904019688269073,
+      "grad_norm": 0.3064846098423004,
+      "learning_rate": 0.0002,
+      "loss": 0.8948,
+      "step": 3490
+    },
+    {
+      "epoch": 0.22969647251845776,
+      "grad_norm": 0.31982484459877014,
+      "learning_rate": 0.0002,
+      "loss": 0.8805,
+      "step": 3500
+    },
+    {
+      "epoch": 0.23035274815422477,
+      "grad_norm": 0.27903324365615845,
+      "learning_rate": 0.0002,
+      "loss": 0.927,
+      "step": 3510
+    },
+    {
+      "epoch": 0.2310090237899918,
+      "grad_norm": 0.35411202907562256,
+      "learning_rate": 0.0002,
+      "loss": 0.9103,
+      "step": 3520
+    },
+    {
+      "epoch": 0.2316652994257588,
+      "grad_norm": 0.2796868681907654,
+      "learning_rate": 0.0002,
+      "loss": 0.9087,
+      "step": 3530
+    },
+    {
+      "epoch": 0.23232157506152584,
+      "grad_norm": 0.3428329825401306,
+      "learning_rate": 0.0002,
+      "loss": 0.9861,
+      "step": 3540
+    },
+    {
+      "epoch": 0.23297785069729288,
+      "grad_norm": 0.30563032627105713,
+      "learning_rate": 0.0002,
+      "loss": 0.9025,
+      "step": 3550
+    },
+    {
+      "epoch": 0.23363412633305988,
+      "grad_norm": 0.2954406142234802,
+      "learning_rate": 0.0002,
+      "loss": 0.9132,
+      "step": 3560
+    },
+    {
+      "epoch": 0.23429040196882692,
+      "grad_norm": 0.3328028917312622,
+      "learning_rate": 0.0002,
+      "loss": 0.9444,
+      "step": 3570
+    },
+    {
+      "epoch": 0.23494667760459392,
+      "grad_norm": 0.32020696997642517,
+      "learning_rate": 0.0002,
+      "loss": 0.9161,
+      "step": 3580
+    },
+    {
+      "epoch": 0.23560295324036096,
+      "grad_norm": 0.2774750292301178,
+      "learning_rate": 0.0002,
+      "loss": 0.9193,
+      "step": 3590
+    },
+    {
+      "epoch": 0.23625922887612796,
+      "grad_norm": 0.24560679495334625,
+      "learning_rate": 0.0002,
+      "loss": 0.8935,
+      "step": 3600
+    },
+    {
+      "epoch": 0.236915504511895,
+      "grad_norm": 0.3278765082359314,
+      "learning_rate": 0.0002,
+      "loss": 0.9335,
+      "step": 3610
+    },
+    {
+      "epoch": 0.23757178014766203,
+      "grad_norm": 0.4417719542980194,
+      "learning_rate": 0.0002,
+      "loss": 0.8958,
+      "step": 3620
+    },
+    {
+      "epoch": 0.23822805578342904,
+      "grad_norm": 0.3339618742465973,
+      "learning_rate": 0.0002,
+      "loss": 0.9683,
+      "step": 3630
+    },
+    {
+      "epoch": 0.23888433141919607,
+      "grad_norm": 0.2925402522087097,
+      "learning_rate": 0.0002,
+      "loss": 0.9536,
+      "step": 3640
+    },
+    {
+      "epoch": 0.23954060705496308,
+      "grad_norm": 0.3060242831707001,
+      "learning_rate": 0.0002,
+      "loss": 0.9695,
+      "step": 3650
+    },
+    {
+      "epoch": 0.2401968826907301,
+      "grad_norm": 0.3095077872276306,
+      "learning_rate": 0.0002,
+      "loss": 0.9573,
+      "step": 3660
+    },
+    {
+      "epoch": 0.24085315832649712,
+      "grad_norm": 0.31151828169822693,
+      "learning_rate": 0.0002,
+      "loss": 0.9023,
+      "step": 3670
+    },
+    {
+      "epoch": 0.24150943396226415,
+      "grad_norm": 0.293610543012619,
+      "learning_rate": 0.0002,
+      "loss": 0.9642,
+      "step": 3680
+    },
+    {
+      "epoch": 0.2421657095980312,
+      "grad_norm": 0.28868040442466736,
+      "learning_rate": 0.0002,
+      "loss": 0.9449,
+      "step": 3690
+    },
+    {
+      "epoch": 0.2428219852337982,
+      "grad_norm": 0.3123609721660614,
+      "learning_rate": 0.0002,
+      "loss": 0.9258,
+      "step": 3700
+    },
+    {
+      "epoch": 0.24347826086956523,
+      "grad_norm": 0.29155978560447693,
+      "learning_rate": 0.0002,
+      "loss": 0.9013,
+      "step": 3710
+    },
+    {
+      "epoch": 0.24413453650533223,
+      "grad_norm": 0.36799871921539307,
+      "learning_rate": 0.0002,
+      "loss": 0.8977,
+      "step": 3720
+    },
+    {
+      "epoch": 0.24479081214109927,
+      "grad_norm": 0.30915290117263794,
+      "learning_rate": 0.0002,
+      "loss": 1.0157,
+      "step": 3730
+    },
+    {
+      "epoch": 0.24544708777686627,
+      "grad_norm": 0.2889885902404785,
+      "learning_rate": 0.0002,
+      "loss": 0.9137,
+      "step": 3740
+    },
+    {
+      "epoch": 0.2461033634126333,
+      "grad_norm": 0.30213671922683716,
+      "learning_rate": 0.0002,
+      "loss": 0.9279,
+      "step": 3750
+    },
+    {
+      "epoch": 0.24675963904840031,
+      "grad_norm": 0.3242565393447876,
+      "learning_rate": 0.0002,
+      "loss": 0.9202,
+      "step": 3760
+    },
+    {
+      "epoch": 0.24741591468416735,
+      "grad_norm": 0.27826303243637085,
+      "learning_rate": 0.0002,
+      "loss": 0.9323,
+      "step": 3770
+    },
+    {
+      "epoch": 0.24807219031993438,
+      "grad_norm": 0.2813768982887268,
+      "learning_rate": 0.0002,
+      "loss": 0.933,
+      "step": 3780
+    },
+    {
+      "epoch": 0.2487284659557014,
+      "grad_norm": 0.34054139256477356,
+      "learning_rate": 0.0002,
+      "loss": 0.9046,
+      "step": 3790
+    },
+    {
+      "epoch": 0.24938474159146842,
+      "grad_norm": 0.2804257869720459,
+      "learning_rate": 0.0002,
+      "loss": 0.9252,
+      "step": 3800
+    },
+    {
+      "epoch": 0.25004101722723543,
+      "grad_norm": 0.2920171618461609,
+      "learning_rate": 0.0002,
+      "loss": 0.8305,
+      "step": 3810
+    },
+    {
+      "epoch": 0.25069729286300246,
+      "grad_norm": 0.2967351973056793,
+      "learning_rate": 0.0002,
+      "loss": 0.883,
+      "step": 3820
+    },
+    {
+      "epoch": 0.2513535684987695,
+      "grad_norm": 0.27417901158332825,
+      "learning_rate": 0.0002,
+      "loss": 0.8767,
+      "step": 3830
+    },
+    {
+      "epoch": 0.25200984413453653,
+      "grad_norm": 0.30916500091552734,
+      "learning_rate": 0.0002,
+      "loss": 0.9459,
+      "step": 3840
+    },
+    {
+      "epoch": 0.2526661197703035,
+      "grad_norm": 0.3281477391719818,
+      "learning_rate": 0.0002,
+      "loss": 0.9774,
+      "step": 3850
+    },
+    {
+      "epoch": 0.25332239540607054,
+      "grad_norm": 0.3417379856109619,
+      "learning_rate": 0.0002,
+      "loss": 0.8779,
+      "step": 3860
+    },
+    {
+      "epoch": 0.2539786710418376,
+      "grad_norm": 0.3119784891605377,
+      "learning_rate": 0.0002,
+      "loss": 0.9155,
+      "step": 3870
+    },
+    {
+      "epoch": 0.2546349466776046,
+      "grad_norm": 0.3053247332572937,
+      "learning_rate": 0.0002,
+      "loss": 0.9466,
+      "step": 3880
+    },
+    {
+      "epoch": 0.2552912223133716,
+      "grad_norm": 0.3155108094215393,
+      "learning_rate": 0.0002,
+      "loss": 0.8882,
+      "step": 3890
+    },
+    {
+      "epoch": 0.2559474979491386,
+      "grad_norm": 0.3240940272808075,
+      "learning_rate": 0.0002,
+      "loss": 0.9464,
+      "step": 3900
+    },
+    {
+      "epoch": 0.25660377358490566,
+      "grad_norm": 0.32433149218559265,
+      "learning_rate": 0.0002,
+      "loss": 0.9765,
+      "step": 3910
+    },
+    {
+      "epoch": 0.2572600492206727,
+      "grad_norm": 0.2982637286186218,
+      "learning_rate": 0.0002,
+      "loss": 0.9714,
+      "step": 3920
+    },
+    {
+      "epoch": 0.2579163248564397,
+      "grad_norm": 0.34742113947868347,
+      "learning_rate": 0.0002,
+      "loss": 0.8758,
+      "step": 3930
+    },
+    {
+      "epoch": 0.2585726004922067,
+      "grad_norm": 0.3415607213973999,
+      "learning_rate": 0.0002,
+      "loss": 0.9335,
+      "step": 3940
+    },
+    {
+      "epoch": 0.25922887612797374,
+      "grad_norm": 0.2852018475532532,
+      "learning_rate": 0.0002,
+      "loss": 0.9016,
+      "step": 3950
+    },
+    {
+      "epoch": 0.2598851517637408,
+      "grad_norm": 0.2924928665161133,
+      "learning_rate": 0.0002,
+      "loss": 0.9972,
+      "step": 3960
+    },
+    {
+      "epoch": 0.2605414273995078,
+      "grad_norm": 0.29387834668159485,
+      "learning_rate": 0.0002,
+      "loss": 0.9524,
+      "step": 3970
+    },
+    {
+      "epoch": 0.26119770303527484,
+      "grad_norm": 0.32400810718536377,
+      "learning_rate": 0.0002,
+      "loss": 0.9197,
+      "step": 3980
+    },
+    {
+      "epoch": 0.2618539786710418,
+      "grad_norm": 0.35629674792289734,
+      "learning_rate": 0.0002,
+      "loss": 0.9886,
+      "step": 3990
+    },
+    {
+      "epoch": 0.26251025430680885,
+      "grad_norm": 0.40292513370513916,
+      "learning_rate": 0.0002,
+      "loss": 0.9002,
+      "step": 4000
+    },
+    {
+      "epoch": 0.2631665299425759,
+      "grad_norm": 0.3452664613723755,
+      "learning_rate": 0.0002,
+      "loss": 0.9561,
+      "step": 4010
+    },
+    {
+      "epoch": 0.2638228055783429,
+      "grad_norm": 0.2956177592277527,
+      "learning_rate": 0.0002,
+      "loss": 0.9231,
+      "step": 4020
+    },
+    {
+      "epoch": 0.2644790812141099,
+      "grad_norm": 0.3159438669681549,
+      "learning_rate": 0.0002,
+      "loss": 0.9269,
+      "step": 4030
+    },
+    {
+      "epoch": 0.26513535684987694,
+      "grad_norm": 0.25804826617240906,
+      "learning_rate": 0.0002,
+      "loss": 0.9682,
+      "step": 4040
+    },
+    {
+      "epoch": 0.26579163248564397,
+      "grad_norm": 0.3446296751499176,
+      "learning_rate": 0.0002,
+      "loss": 0.9681,
+      "step": 4050
+    },
+    {
+      "epoch": 0.266447908121411,
+      "grad_norm": 0.32568585872650146,
+      "learning_rate": 0.0002,
+      "loss": 0.9451,
+      "step": 4060
+    },
+    {
+      "epoch": 0.26710418375717804,
+      "grad_norm": 0.2872511148452759,
+      "learning_rate": 0.0002,
+      "loss": 0.907,
+      "step": 4070
+    },
+    {
+      "epoch": 0.267760459392945,
+      "grad_norm": 0.327495276927948,
+      "learning_rate": 0.0002,
+      "loss": 0.9365,
+      "step": 4080
+    },
+    {
+      "epoch": 0.26841673502871205,
+      "grad_norm": 0.40853601694107056,
+      "learning_rate": 0.0002,
+      "loss": 0.9778,
+      "step": 4090
+    },
+    {
+      "epoch": 0.2690730106644791,
+      "grad_norm": 0.32097089290618896,
+      "learning_rate": 0.0002,
+      "loss": 0.9232,
+      "step": 4100
+    },
+    {
+      "epoch": 0.2697292863002461,
+      "grad_norm": 0.32956209778785706,
+      "learning_rate": 0.0002,
+      "loss": 0.9622,
+      "step": 4110
+    },
+    {
+      "epoch": 0.27038556193601315,
+      "grad_norm": 0.30057013034820557,
+      "learning_rate": 0.0002,
+      "loss": 0.911,
+      "step": 4120
+    },
+    {
+      "epoch": 0.27104183757178013,
+      "grad_norm": 0.30630001425743103,
+      "learning_rate": 0.0002,
+      "loss": 0.9045,
+      "step": 4130
+    },
+    {
+      "epoch": 0.27169811320754716,
+      "grad_norm": 0.31977957487106323,
+      "learning_rate": 0.0002,
+      "loss": 0.8914,
+      "step": 4140
+    },
+    {
+      "epoch": 0.2723543888433142,
+      "grad_norm": 0.29783955216407776,
+      "learning_rate": 0.0002,
+      "loss": 0.9619,
+      "step": 4150
+    },
+    {
+      "epoch": 0.27301066447908123,
+      "grad_norm": 0.37371826171875,
+      "learning_rate": 0.0002,
+      "loss": 1.0299,
+      "step": 4160
+    },
+    {
+      "epoch": 0.2736669401148482,
+      "grad_norm": 0.27656471729278564,
+      "learning_rate": 0.0002,
+      "loss": 0.9072,
+      "step": 4170
+    },
+    {
+      "epoch": 0.27432321575061525,
+      "grad_norm": 0.29685157537460327,
+      "learning_rate": 0.0002,
+      "loss": 0.873,
+      "step": 4180
+    },
+    {
+      "epoch": 0.2749794913863823,
+      "grad_norm": 0.3376981317996979,
+      "learning_rate": 0.0002,
+      "loss": 1.0057,
+      "step": 4190
+    },
+    {
+      "epoch": 0.2756357670221493,
+      "grad_norm": 0.2933194041252136,
+      "learning_rate": 0.0002,
+      "loss": 0.9671,
+      "step": 4200
+    },
+    {
+      "epoch": 0.27629204265791635,
+      "grad_norm": 0.27594974637031555,
+      "learning_rate": 0.0002,
+      "loss": 0.9359,
+      "step": 4210
+    },
+    {
+      "epoch": 0.2769483182936833,
+      "grad_norm": 0.29267510771751404,
+      "learning_rate": 0.0002,
+      "loss": 0.9336,
+      "step": 4220
+    },
+    {
+      "epoch": 0.27760459392945036,
+      "grad_norm": 0.3266076445579529,
+      "learning_rate": 0.0002,
+      "loss": 0.9241,
+      "step": 4230
+    },
+    {
+      "epoch": 0.2782608695652174,
+      "grad_norm": 0.3626921474933624,
+      "learning_rate": 0.0002,
+      "loss": 0.933,
+      "step": 4240
+    },
+    {
+      "epoch": 0.27891714520098443,
+      "grad_norm": 0.3043886423110962,
+      "learning_rate": 0.0002,
+      "loss": 0.9533,
+      "step": 4250
+    },
+    {
+      "epoch": 0.27957342083675146,
+      "grad_norm": 0.28676939010620117,
+      "learning_rate": 0.0002,
+      "loss": 0.902,
+      "step": 4260
+    },
+    {
+      "epoch": 0.28022969647251844,
+      "grad_norm": 0.32501107454299927,
+      "learning_rate": 0.0002,
+      "loss": 0.9565,
+      "step": 4270
+    },
+    {
+      "epoch": 0.2808859721082855,
+      "grad_norm": 0.3533550202846527,
+      "learning_rate": 0.0002,
+      "loss": 1.0009,
+      "step": 4280
+    },
+    {
+      "epoch": 0.2815422477440525,
+      "grad_norm": 0.28312650322914124,
+      "learning_rate": 0.0002,
+      "loss": 0.9128,
+      "step": 4290
+    },
+    {
+      "epoch": 0.28219852337981954,
+      "grad_norm": 0.2910906672477722,
+      "learning_rate": 0.0002,
+      "loss": 0.9626,
+      "step": 4300
+    },
+    {
+      "epoch": 0.2828547990155865,
+      "grad_norm": 0.27612248063087463,
+      "learning_rate": 0.0002,
+      "loss": 0.8803,
+      "step": 4310
+    },
+    {
+      "epoch": 0.28351107465135356,
+      "grad_norm": 0.416061133146286,
+      "learning_rate": 0.0002,
+      "loss": 0.9451,
+      "step": 4320
+    },
+    {
+      "epoch": 0.2841673502871206,
+      "grad_norm": 0.36900338530540466,
+      "learning_rate": 0.0002,
+      "loss": 0.9197,
+      "step": 4330
+    },
+    {
+      "epoch": 0.2848236259228876,
+      "grad_norm": 0.32224977016448975,
+      "learning_rate": 0.0002,
+      "loss": 0.9623,
+      "step": 4340
+    },
+    {
+      "epoch": 0.28547990155865466,
+      "grad_norm": 0.2805930972099304,
+      "learning_rate": 0.0002,
+      "loss": 0.9264,
+      "step": 4350
+    },
+    {
+      "epoch": 0.28613617719442164,
+      "grad_norm": 0.286539226770401,
+      "learning_rate": 0.0002,
+      "loss": 0.9486,
+      "step": 4360
+    },
+    {
+      "epoch": 0.28679245283018867,
+      "grad_norm": 0.3326348066329956,
+      "learning_rate": 0.0002,
+      "loss": 0.9225,
+      "step": 4370
+    },
+    {
+      "epoch": 0.2874487284659557,
+      "grad_norm": 0.29906603693962097,
+      "learning_rate": 0.0002,
+      "loss": 0.8774,
+      "step": 4380
+    },
+    {
+      "epoch": 0.28810500410172274,
+      "grad_norm": 0.33906206488609314,
+      "learning_rate": 0.0002,
+      "loss": 0.9582,
+      "step": 4390
+    },
+    {
+      "epoch": 0.2887612797374898,
+      "grad_norm": 0.36083030700683594,
+      "learning_rate": 0.0002,
+      "loss": 0.8987,
+      "step": 4400
+    },
+    {
+      "epoch": 0.28941755537325675,
+      "grad_norm": 0.30945461988449097,
+      "learning_rate": 0.0002,
+      "loss": 0.9189,
+      "step": 4410
+    },
+    {
+      "epoch": 0.2900738310090238,
+      "grad_norm": 0.29673129320144653,
+      "learning_rate": 0.0002,
+      "loss": 0.9454,
+      "step": 4420
+    },
+    {
+      "epoch": 0.2907301066447908,
+      "grad_norm": 0.29949837923049927,
+      "learning_rate": 0.0002,
+      "loss": 0.9454,
+      "step": 4430
+    },
+    {
+      "epoch": 0.29138638228055785,
+      "grad_norm": 0.29319390654563904,
+      "learning_rate": 0.0002,
+      "loss": 0.9026,
+      "step": 4440
+    },
+    {
+      "epoch": 0.29204265791632483,
+      "grad_norm": 0.38954150676727295,
+      "learning_rate": 0.0002,
+      "loss": 0.9318,
+      "step": 4450
+    },
+    {
+      "epoch": 0.29269893355209187,
+      "grad_norm": 0.32702240347862244,
+      "learning_rate": 0.0002,
+      "loss": 1.0558,
+      "step": 4460
+    },
+    {
+      "epoch": 0.2933552091878589,
+      "grad_norm": 0.3480510413646698,
+      "learning_rate": 0.0002,
+      "loss": 0.9684,
+      "step": 4470
+    },
+    {
+      "epoch": 0.29401148482362593,
+      "grad_norm": 0.34290337562561035,
+      "learning_rate": 0.0002,
+      "loss": 0.9267,
+      "step": 4480
+    },
+    {
+      "epoch": 0.29466776045939297,
+      "grad_norm": 0.3169049620628357,
+      "learning_rate": 0.0002,
+      "loss": 0.9561,
+      "step": 4490
+    },
+    {
+      "epoch": 0.29532403609515995,
+      "grad_norm": 0.3368853032588959,
+      "learning_rate": 0.0002,
+      "loss": 0.9386,
+      "step": 4500
+    },
+    {
+      "epoch": 0.295980311730927,
+      "grad_norm": 0.30214768648147583,
+      "learning_rate": 0.0002,
+      "loss": 0.8803,
+      "step": 4510
+    },
+    {
+      "epoch": 0.296636587366694,
+      "grad_norm": 0.33051690459251404,
+      "learning_rate": 0.0002,
+      "loss": 0.9485,
+      "step": 4520
+    },
+    {
+      "epoch": 0.29729286300246105,
+      "grad_norm": 0.37137898802757263,
+      "learning_rate": 0.0002,
+      "loss": 0.9498,
+      "step": 4530
+    },
+    {
+      "epoch": 0.29794913863822803,
+      "grad_norm": 0.3780321180820465,
+      "learning_rate": 0.0002,
+      "loss": 0.9518,
+      "step": 4540
+    },
+    {
+      "epoch": 0.29860541427399506,
+      "grad_norm": 0.3048851788043976,
+      "learning_rate": 0.0002,
+      "loss": 0.8912,
+      "step": 4550
+    },
+    {
+      "epoch": 0.2992616899097621,
+      "grad_norm": 0.36871910095214844,
+      "learning_rate": 0.0002,
+      "loss": 1.0054,
+      "step": 4560
+    },
+    {
+      "epoch": 0.29991796554552913,
+      "grad_norm": 0.3335227966308594,
+      "learning_rate": 0.0002,
+      "loss": 0.9172,
+      "step": 4570
+    },
+    {
+      "epoch": 0.30057424118129616,
+      "grad_norm": 0.3256683349609375,
+      "learning_rate": 0.0002,
+      "loss": 0.9368,
+      "step": 4580
+    },
+    {
+      "epoch": 0.30123051681706314,
+      "grad_norm": 0.34209194779396057,
+      "learning_rate": 0.0002,
+      "loss": 0.9641,
+      "step": 4590
+    },
+    {
+      "epoch": 0.3018867924528302,
+      "grad_norm": 0.278749018907547,
+      "learning_rate": 0.0002,
+      "loss": 0.9128,
+      "step": 4600
+    },
+    {
+      "epoch": 0.3025430680885972,
+      "grad_norm": 0.28081533312797546,
+      "learning_rate": 0.0002,
+      "loss": 0.931,
+      "step": 4610
+    },
+    {
+      "epoch": 0.30319934372436425,
+      "grad_norm": 0.35687389969825745,
+      "learning_rate": 0.0002,
+      "loss": 0.9084,
+      "step": 4620
+    },
+    {
+      "epoch": 0.3038556193601313,
+      "grad_norm": 0.367288738489151,
+      "learning_rate": 0.0002,
+      "loss": 0.8927,
+      "step": 4630
+    },
+    {
+      "epoch": 0.30451189499589826,
+      "grad_norm": 0.40433239936828613,
+      "learning_rate": 0.0002,
+      "loss": 0.9175,
+      "step": 4640
+    },
+    {
+      "epoch": 0.3051681706316653,
+      "grad_norm": 0.31489279866218567,
+      "learning_rate": 0.0002,
+      "loss": 0.9115,
+      "step": 4650
+    },
+    {
+      "epoch": 0.3058244462674323,
+      "grad_norm": 0.30823174118995667,
+      "learning_rate": 0.0002,
+      "loss": 0.9843,
+      "step": 4660
+    },
+    {
+      "epoch": 0.30648072190319936,
+      "grad_norm": 0.3204546570777893,
+      "learning_rate": 0.0002,
+      "loss": 0.964,
+      "step": 4670
+    },
+    {
+      "epoch": 0.30713699753896634,
+      "grad_norm": 0.3188243508338928,
+      "learning_rate": 0.0002,
+      "loss": 0.9754,
+      "step": 4680
+    },
+    {
+      "epoch": 0.3077932731747334,
+      "grad_norm": 0.31931981444358826,
+      "learning_rate": 0.0002,
+      "loss": 0.8923,
+      "step": 4690
+    },
+    {
+      "epoch": 0.3084495488105004,
+      "grad_norm": 0.33908605575561523,
+      "learning_rate": 0.0002,
+      "loss": 0.9113,
+      "step": 4700
+    },
+    {
+      "epoch": 0.30910582444626744,
+      "grad_norm": 0.3362937569618225,
+      "learning_rate": 0.0002,
+      "loss": 0.9552,
+      "step": 4710
+    },
+    {
+      "epoch": 0.3097621000820345,
+      "grad_norm": 0.3304995000362396,
+      "learning_rate": 0.0002,
+      "loss": 0.9695,
+      "step": 4720
+    },
+    {
+      "epoch": 0.31041837571780145,
+      "grad_norm": 0.39165404438972473,
+      "learning_rate": 0.0002,
+      "loss": 0.9588,
+      "step": 4730
+    },
+    {
+      "epoch": 0.3110746513535685,
+      "grad_norm": 0.29097774624824524,
+      "learning_rate": 0.0002,
+      "loss": 0.9583,
+      "step": 4740
+    },
+    {
+      "epoch": 0.3117309269893355,
+      "grad_norm": 0.289065420627594,
+      "learning_rate": 0.0002,
+      "loss": 0.9224,
+      "step": 4750
+    },
+    {
+      "epoch": 0.31238720262510256,
+      "grad_norm": 0.32188501954078674,
+      "learning_rate": 0.0002,
+      "loss": 0.8905,
+      "step": 4760
+    },
+    {
+      "epoch": 0.3130434782608696,
+      "grad_norm": 0.2738335430622101,
+      "learning_rate": 0.0002,
+      "loss": 0.951,
+      "step": 4770
+    },
+    {
+      "epoch": 0.31369975389663657,
+      "grad_norm": 0.2811134457588196,
+      "learning_rate": 0.0002,
+      "loss": 0.9662,
+      "step": 4780
+    },
+    {
+      "epoch": 0.3143560295324036,
+      "grad_norm": 0.2986338138580322,
+      "learning_rate": 0.0002,
+      "loss": 0.9372,
+      "step": 4790
+    },
+    {
+      "epoch": 0.31501230516817064,
+      "grad_norm": 0.30152231454849243,
+      "learning_rate": 0.0002,
+      "loss": 0.9016,
+      "step": 4800
+    },
+    {
+      "epoch": 0.31566858080393767,
+      "grad_norm": 0.37527933716773987,
+      "learning_rate": 0.0002,
+      "loss": 0.9902,
+      "step": 4810
+    },
+    {
+      "epoch": 0.31632485643970465,
+      "grad_norm": 0.4206887483596802,
+      "learning_rate": 0.0002,
+      "loss": 0.9786,
+      "step": 4820
+    },
+    {
+      "epoch": 0.3169811320754717,
+      "grad_norm": 0.3060953617095947,
+      "learning_rate": 0.0002,
+      "loss": 0.9265,
+      "step": 4830
+    },
+    {
+      "epoch": 0.3176374077112387,
+      "grad_norm": 0.28399568796157837,
+      "learning_rate": 0.0002,
+      "loss": 0.874,
+      "step": 4840
+    },
+    {
+      "epoch": 0.31829368334700575,
+      "grad_norm": 0.30498018860816956,
+      "learning_rate": 0.0002,
+      "loss": 0.9287,
+      "step": 4850
+    },
+    {
+      "epoch": 0.3189499589827728,
+      "grad_norm": 0.3168811798095703,
+      "learning_rate": 0.0002,
+      "loss": 0.9682,
+      "step": 4860
+    },
+    {
+      "epoch": 0.31960623461853976,
+      "grad_norm": 0.33051279187202454,
+      "learning_rate": 0.0002,
+      "loss": 0.9916,
+      "step": 4870
+    },
+    {
+      "epoch": 0.3202625102543068,
+      "grad_norm": 0.34563300013542175,
+      "learning_rate": 0.0002,
+      "loss": 0.9427,
+      "step": 4880
+    },
+    {
+      "epoch": 0.32091878589007383,
+      "grad_norm": 0.28437602519989014,
+      "learning_rate": 0.0002,
+      "loss": 0.9018,
+      "step": 4890
+    },
+    {
+      "epoch": 0.32157506152584087,
+      "grad_norm": 0.3883301913738251,
+      "learning_rate": 0.0002,
+      "loss": 0.9357,
+      "step": 4900
+    },
+    {
+      "epoch": 0.3222313371616079,
+      "grad_norm": 0.2933879792690277,
+      "learning_rate": 0.0002,
+      "loss": 0.9034,
+      "step": 4910
+    },
+    {
+      "epoch": 0.3228876127973749,
+      "grad_norm": 0.3490493595600128,
+      "learning_rate": 0.0002,
+      "loss": 0.9332,
+      "step": 4920
+    },
+    {
+      "epoch": 0.3235438884331419,
+      "grad_norm": 0.29370415210723877,
+      "learning_rate": 0.0002,
+      "loss": 0.9424,
+      "step": 4930
+    },
+    {
+      "epoch": 0.32420016406890895,
+      "grad_norm": 0.32653453946113586,
+      "learning_rate": 0.0002,
+      "loss": 0.9751,
+      "step": 4940
+    },
+    {
+      "epoch": 0.324856439704676,
+      "grad_norm": 0.32540783286094666,
+      "learning_rate": 0.0002,
+      "loss": 0.9278,
+      "step": 4950
+    },
+    {
+      "epoch": 0.32551271534044296,
+      "grad_norm": 0.37576451897621155,
+      "learning_rate": 0.0002,
+      "loss": 0.9309,
+      "step": 4960
+    },
+    {
+      "epoch": 0.32616899097621,
+      "grad_norm": 0.27892962098121643,
+      "learning_rate": 0.0002,
+      "loss": 0.9612,
+      "step": 4970
+    },
+    {
+      "epoch": 0.32682526661197703,
+      "grad_norm": 0.38215309381484985,
+      "learning_rate": 0.0002,
+      "loss": 0.9765,
+      "step": 4980
+    },
+    {
+      "epoch": 0.32748154224774406,
+      "grad_norm": 0.4062992334365845,
+      "learning_rate": 0.0002,
+      "loss": 0.9226,
+      "step": 4990
+    },
+    {
+      "epoch": 0.3281378178835111,
+      "grad_norm": 0.3053376376628876,
+      "learning_rate": 0.0002,
+      "loss": 0.9157,
+      "step": 5000
+    },
+    {
+      "epoch": 0.3287940935192781,
+      "grad_norm": 0.3228836953639984,
+      "learning_rate": 0.0002,
+      "loss": 0.9117,
+      "step": 5010
+    },
+    {
+      "epoch": 0.3294503691550451,
+      "grad_norm": 0.27714234590530396,
+      "learning_rate": 0.0002,
+      "loss": 0.8644,
+      "step": 5020
+    },
+    {
+      "epoch": 0.33010664479081214,
+      "grad_norm": 0.3127416968345642,
+      "learning_rate": 0.0002,
+      "loss": 0.9034,
+      "step": 5030
+    },
+    {
+      "epoch": 0.3307629204265792,
+      "grad_norm": 0.3653234839439392,
+      "learning_rate": 0.0002,
+      "loss": 0.89,
+      "step": 5040
+    },
+    {
+      "epoch": 0.3314191960623462,
+      "grad_norm": 0.33517029881477356,
+      "learning_rate": 0.0002,
+      "loss": 0.9313,
+      "step": 5050
+    },
+    {
+      "epoch": 0.3320754716981132,
+      "grad_norm": 0.3412803113460541,
+      "learning_rate": 0.0002,
+      "loss": 0.9219,
+      "step": 5060
+    },
+    {
+      "epoch": 0.3327317473338802,
+      "grad_norm": 0.32958096265792847,
+      "learning_rate": 0.0002,
+      "loss": 0.9358,
+      "step": 5070
+    },
+    {
+      "epoch": 0.33338802296964726,
+      "grad_norm": 0.2991558015346527,
+      "learning_rate": 0.0002,
+      "loss": 0.9479,
+      "step": 5080
+    },
+    {
+      "epoch": 0.3340442986054143,
+      "grad_norm": 0.35870906710624695,
+      "learning_rate": 0.0002,
+      "loss": 0.9754,
+      "step": 5090
+    },
+    {
+      "epoch": 0.33470057424118127,
+      "grad_norm": 0.26751458644866943,
+      "learning_rate": 0.0002,
+      "loss": 0.9183,
+      "step": 5100
+    },
+    {
+      "epoch": 0.3353568498769483,
+      "grad_norm": 0.35588568449020386,
+      "learning_rate": 0.0002,
+      "loss": 0.9676,
+      "step": 5110
+    },
+    {
+      "epoch": 0.33601312551271534,
+      "grad_norm": 0.41509315371513367,
+      "learning_rate": 0.0002,
+      "loss": 0.8964,
+      "step": 5120
+    },
+    {
+      "epoch": 0.3366694011484824,
+      "grad_norm": 0.37643107771873474,
+      "learning_rate": 0.0002,
+      "loss": 0.9864,
+      "step": 5130
+    },
+    {
+      "epoch": 0.3373256767842494,
+      "grad_norm": 0.3062657117843628,
+      "learning_rate": 0.0002,
+      "loss": 0.9367,
+      "step": 5140
+    },
+    {
+      "epoch": 0.3379819524200164,
+      "grad_norm": 0.2778759300708771,
+      "learning_rate": 0.0002,
+      "loss": 0.9167,
+      "step": 5150
+    },
+    {
+      "epoch": 0.3386382280557834,
+      "grad_norm": 0.40632501244544983,
+      "learning_rate": 0.0002,
+      "loss": 0.9787,
+      "step": 5160
+    },
+    {
+      "epoch": 0.33929450369155045,
+      "grad_norm": 0.2996899485588074,
+      "learning_rate": 0.0002,
+      "loss": 0.8941,
+      "step": 5170
+    },
+    {
+      "epoch": 0.3399507793273175,
+      "grad_norm": 0.2869918942451477,
+      "learning_rate": 0.0002,
+      "loss": 0.9611,
+      "step": 5180
+    },
+    {
+      "epoch": 0.3406070549630845,
+      "grad_norm": 0.3303343951702118,
+      "learning_rate": 0.0002,
+      "loss": 0.9201,
+      "step": 5190
+    },
+    {
+      "epoch": 0.3412633305988515,
+      "grad_norm": 0.3582284450531006,
+      "learning_rate": 0.0002,
+      "loss": 0.9162,
+      "step": 5200
+    },
+    {
+      "epoch": 0.34191960623461853,
+      "grad_norm": 0.3600800037384033,
+      "learning_rate": 0.0002,
+      "loss": 0.9441,
+      "step": 5210
+    },
+    {
+      "epoch": 0.34257588187038557,
+      "grad_norm": 0.3147357106208801,
+      "learning_rate": 0.0002,
+      "loss": 0.9061,
+      "step": 5220
+    },
+    {
+      "epoch": 0.3432321575061526,
+      "grad_norm": 0.3327115774154663,
+      "learning_rate": 0.0002,
+      "loss": 0.9189,
+      "step": 5230
+    },
+    {
+      "epoch": 0.3438884331419196,
+      "grad_norm": 0.3514555096626282,
+      "learning_rate": 0.0002,
+      "loss": 0.9606,
+      "step": 5240
+    },
+    {
+      "epoch": 0.3445447087776866,
+      "grad_norm": 0.35748785734176636,
+      "learning_rate": 0.0002,
+      "loss": 0.8944,
+      "step": 5250
+    },
+    {
+      "epoch": 0.34520098441345365,
+      "grad_norm": 0.3179738223552704,
+      "learning_rate": 0.0002,
+      "loss": 0.9228,
+      "step": 5260
+    },
+    {
+      "epoch": 0.3458572600492207,
+      "grad_norm": 0.3399673104286194,
+      "learning_rate": 0.0002,
+      "loss": 0.9313,
+      "step": 5270
+    },
+    {
+      "epoch": 0.3465135356849877,
+      "grad_norm": 0.32293835282325745,
+      "learning_rate": 0.0002,
+      "loss": 0.9379,
+      "step": 5280
+    },
+    {
+      "epoch": 0.3471698113207547,
+      "grad_norm": 0.3028234839439392,
+      "learning_rate": 0.0002,
+      "loss": 0.9066,
+      "step": 5290
+    },
+    {
+      "epoch": 0.34782608695652173,
+      "grad_norm": 0.3056369423866272,
+      "learning_rate": 0.0002,
+      "loss": 0.929,
+      "step": 5300
+    },
+    {
+      "epoch": 0.34848236259228876,
+      "grad_norm": 0.36918318271636963,
+      "learning_rate": 0.0002,
+      "loss": 0.9597,
+      "step": 5310
+    },
+    {
+      "epoch": 0.3491386382280558,
+      "grad_norm": 0.2661174535751343,
+      "learning_rate": 0.0002,
+      "loss": 0.9365,
+      "step": 5320
+    },
+    {
+      "epoch": 0.34979491386382283,
+      "grad_norm": 0.294240266084671,
+      "learning_rate": 0.0002,
+      "loss": 0.8832,
+      "step": 5330
+    },
+    {
+      "epoch": 0.3504511894995898,
+      "grad_norm": 0.3509284555912018,
+      "learning_rate": 0.0002,
+      "loss": 0.9135,
+      "step": 5340
+    },
+    {
+      "epoch": 0.35110746513535684,
+      "grad_norm": 0.331851989030838,
+      "learning_rate": 0.0002,
+      "loss": 0.8919,
+      "step": 5350
+    },
+    {
+      "epoch": 0.3517637407711239,
+      "grad_norm": 0.39378881454467773,
+      "learning_rate": 0.0002,
+      "loss": 0.9357,
+      "step": 5360
+    },
+    {
+      "epoch": 0.3524200164068909,
+      "grad_norm": 0.3014044761657715,
+      "learning_rate": 0.0002,
+      "loss": 0.8844,
+      "step": 5370
+    },
+    {
+      "epoch": 0.3530762920426579,
+      "grad_norm": 0.3350718021392822,
+      "learning_rate": 0.0002,
+      "loss": 0.9234,
+      "step": 5380
+    },
+    {
+      "epoch": 0.3537325676784249,
+      "grad_norm": 0.29478874802589417,
+      "learning_rate": 0.0002,
+      "loss": 0.9119,
+      "step": 5390
+    },
+    {
+      "epoch": 0.35438884331419196,
+      "grad_norm": 0.3219947814941406,
+      "learning_rate": 0.0002,
+      "loss": 0.8781,
+      "step": 5400
+    },
+    {
+      "epoch": 0.355045118949959,
+      "grad_norm": 0.31258803606033325,
+      "learning_rate": 0.0002,
+      "loss": 0.9527,
+      "step": 5410
+    },
+    {
+      "epoch": 0.355701394585726,
+      "grad_norm": 0.2951710820198059,
+      "learning_rate": 0.0002,
+      "loss": 0.9124,
+      "step": 5420
+    },
+    {
+      "epoch": 0.356357670221493,
+      "grad_norm": 0.3726331293582916,
+      "learning_rate": 0.0002,
+      "loss": 0.8997,
+      "step": 5430
+    },
+    {
+      "epoch": 0.35701394585726004,
+      "grad_norm": 0.3571377098560333,
+      "learning_rate": 0.0002,
+      "loss": 0.9288,
+      "step": 5440
+    },
+    {
+      "epoch": 0.3576702214930271,
+      "grad_norm": 0.3004431128501892,
+      "learning_rate": 0.0002,
+      "loss": 0.9538,
+      "step": 5450
+    },
+    {
+      "epoch": 0.3583264971287941,
+      "grad_norm": 0.3258959949016571,
+      "learning_rate": 0.0002,
+      "loss": 0.9466,
+      "step": 5460
+    },
+    {
+      "epoch": 0.35898277276456114,
+      "grad_norm": 0.31265372037887573,
+      "learning_rate": 0.0002,
+      "loss": 0.8999,
+      "step": 5470
+    },
+    {
+      "epoch": 0.3596390484003281,
+      "grad_norm": 0.3120972514152527,
+      "learning_rate": 0.0002,
+      "loss": 0.9438,
+      "step": 5480
+    },
+    {
+      "epoch": 0.36029532403609515,
+      "grad_norm": 0.31449300050735474,
+      "learning_rate": 0.0002,
+      "loss": 0.9541,
+      "step": 5490
+    },
+    {
+      "epoch": 0.3609515996718622,
+      "grad_norm": 0.378205806016922,
+      "learning_rate": 0.0002,
+      "loss": 0.9569,
+      "step": 5500
+    },
+    {
+      "epoch": 0.3616078753076292,
+      "grad_norm": 0.34647420048713684,
+      "learning_rate": 0.0002,
+      "loss": 0.8964,
+      "step": 5510
+    },
+    {
+      "epoch": 0.3622641509433962,
+      "grad_norm": 0.34195807576179504,
+      "learning_rate": 0.0002,
+      "loss": 0.9965,
+      "step": 5520
+    },
+    {
+      "epoch": 0.36292042657916324,
+      "grad_norm": 0.35365694761276245,
+      "learning_rate": 0.0002,
+      "loss": 0.8916,
+      "step": 5530
+    },
+    {
+      "epoch": 0.36357670221493027,
+      "grad_norm": 0.3188548982143402,
+      "learning_rate": 0.0002,
+      "loss": 0.9429,
+      "step": 5540
+    },
+    {
+      "epoch": 0.3642329778506973,
+      "grad_norm": 0.35012900829315186,
+      "learning_rate": 0.0002,
+      "loss": 0.9258,
+      "step": 5550
+    },
+    {
+      "epoch": 0.36488925348646434,
+      "grad_norm": 0.3564446270465851,
+      "learning_rate": 0.0002,
+      "loss": 0.9338,
+      "step": 5560
+    },
+    {
+      "epoch": 0.3655455291222313,
+      "grad_norm": 0.29641905426979065,
+      "learning_rate": 0.0002,
+      "loss": 0.9531,
+      "step": 5570
+    },
+    {
+      "epoch": 0.36620180475799835,
+      "grad_norm": 0.3476453721523285,
+      "learning_rate": 0.0002,
+      "loss": 0.9507,
+      "step": 5580
+    },
+    {
+      "epoch": 0.3668580803937654,
+      "grad_norm": 0.32947802543640137,
+      "learning_rate": 0.0002,
+      "loss": 0.9509,
+      "step": 5590
+    },
+    {
+      "epoch": 0.3675143560295324,
+      "grad_norm": 0.28639012575149536,
+      "learning_rate": 0.0002,
+      "loss": 0.9577,
+      "step": 5600
+    },
+    {
+      "epoch": 0.36817063166529945,
+      "grad_norm": 0.35645192861557007,
+      "learning_rate": 0.0002,
+      "loss": 0.9196,
+      "step": 5610
+    },
+    {
+      "epoch": 0.36882690730106643,
+      "grad_norm": 0.3594934046268463,
+      "learning_rate": 0.0002,
+      "loss": 0.9544,
+      "step": 5620
+    },
+    {
+      "epoch": 0.36948318293683347,
+      "grad_norm": 0.3584282398223877,
+      "learning_rate": 0.0002,
+      "loss": 0.9081,
+      "step": 5630
+    },
+    {
+      "epoch": 0.3701394585726005,
+      "grad_norm": 0.32094448804855347,
+      "learning_rate": 0.0002,
+      "loss": 0.955,
+      "step": 5640
+    },
+    {
+      "epoch": 0.37079573420836753,
+      "grad_norm": 0.3229917287826538,
+      "learning_rate": 0.0002,
+      "loss": 0.9271,
+      "step": 5650
+    },
+    {
+      "epoch": 0.3714520098441345,
+      "grad_norm": 0.3371448218822479,
+      "learning_rate": 0.0002,
+      "loss": 0.933,
+      "step": 5660
+    },
+    {
+      "epoch": 0.37210828547990155,
+      "grad_norm": 0.38639435172080994,
+      "learning_rate": 0.0002,
+      "loss": 0.9189,
+      "step": 5670
+    },
+    {
+      "epoch": 0.3727645611156686,
+      "grad_norm": 0.37710604071617126,
+      "learning_rate": 0.0002,
+      "loss": 0.8987,
+      "step": 5680
+    },
+    {
+      "epoch": 0.3734208367514356,
+      "grad_norm": 0.28154921531677246,
+      "learning_rate": 0.0002,
+      "loss": 0.9133,
+      "step": 5690
+    },
+    {
+      "epoch": 0.37407711238720265,
+      "grad_norm": 0.35519927740097046,
+      "learning_rate": 0.0002,
+      "loss": 0.9104,
+      "step": 5700
+    },
+    {
+      "epoch": 0.3747333880229696,
+      "grad_norm": 0.3271195888519287,
+      "learning_rate": 0.0002,
+      "loss": 0.9187,
+      "step": 5710
+    },
+    {
+      "epoch": 0.37538966365873666,
+      "grad_norm": 0.30836206674575806,
+      "learning_rate": 0.0002,
+      "loss": 1.0015,
+      "step": 5720
+    },
+    {
+      "epoch": 0.3760459392945037,
+      "grad_norm": 0.36397358775138855,
+      "learning_rate": 0.0002,
+      "loss": 0.8888,
+      "step": 5730
+    },
+    {
+      "epoch": 0.37670221493027073,
+      "grad_norm": 0.31649520993232727,
+      "learning_rate": 0.0002,
+      "loss": 0.8962,
+      "step": 5740
+    },
+    {
+      "epoch": 0.37735849056603776,
+      "grad_norm": 0.3790256977081299,
+      "learning_rate": 0.0002,
+      "loss": 0.9667,
+      "step": 5750
+    },
+    {
+      "epoch": 0.37801476620180474,
+      "grad_norm": 0.34175634384155273,
+      "learning_rate": 0.0002,
+      "loss": 0.9548,
+      "step": 5760
+    },
+    {
+      "epoch": 0.3786710418375718,
+      "grad_norm": 0.4169650673866272,
+      "learning_rate": 0.0002,
+      "loss": 0.9031,
+      "step": 5770
+    },
+    {
+      "epoch": 0.3793273174733388,
+      "grad_norm": 0.3118833899497986,
+      "learning_rate": 0.0002,
+      "loss": 0.9279,
+      "step": 5780
+    },
+    {
+      "epoch": 0.37998359310910584,
+      "grad_norm": 0.3474520444869995,
+      "learning_rate": 0.0002,
+      "loss": 0.9204,
+      "step": 5790
+    },
+    {
+      "epoch": 0.3806398687448728,
+      "grad_norm": 0.40819892287254333,
+      "learning_rate": 0.0002,
+      "loss": 0.9845,
+      "step": 5800
+    },
+    {
+      "epoch": 0.38129614438063986,
+      "grad_norm": 0.34402596950531006,
+      "learning_rate": 0.0002,
+      "loss": 0.8863,
+      "step": 5810
+    },
+    {
+      "epoch": 0.3819524200164069,
+      "grad_norm": 0.31899532675743103,
+      "learning_rate": 0.0002,
+      "loss": 0.9356,
+      "step": 5820
+    },
+    {
+      "epoch": 0.3826086956521739,
+      "grad_norm": 0.38860151171684265,
+      "learning_rate": 0.0002,
+      "loss": 0.857,
+      "step": 5830
+    },
+    {
+      "epoch": 0.38326497128794096,
+      "grad_norm": 0.2960244417190552,
+      "learning_rate": 0.0002,
+      "loss": 0.9433,
+      "step": 5840
+    },
+    {
+      "epoch": 0.38392124692370794,
+      "grad_norm": 0.3614438772201538,
+      "learning_rate": 0.0002,
+      "loss": 0.9546,
+      "step": 5850
+    },
+    {
+      "epoch": 0.38457752255947497,
+      "grad_norm": 0.33000093698501587,
+      "learning_rate": 0.0002,
+      "loss": 0.9447,
+      "step": 5860
+    },
+    {
+      "epoch": 0.385233798195242,
+      "grad_norm": 0.3543432354927063,
+      "learning_rate": 0.0002,
+      "loss": 0.964,
+      "step": 5870
+    },
+    {
+      "epoch": 0.38589007383100904,
+      "grad_norm": 0.3281349837779999,
+      "learning_rate": 0.0002,
+      "loss": 0.9329,
+      "step": 5880
+    },
+    {
+      "epoch": 0.3865463494667761,
+      "grad_norm": 0.3775436580181122,
+      "learning_rate": 0.0002,
+      "loss": 0.9463,
+      "step": 5890
+    },
+    {
+      "epoch": 0.38720262510254305,
+      "grad_norm": 0.3037714660167694,
+      "learning_rate": 0.0002,
+      "loss": 0.9035,
+      "step": 5900
+    },
+    {
+      "epoch": 0.3878589007383101,
+      "grad_norm": 0.30885955691337585,
+      "learning_rate": 0.0002,
+      "loss": 0.8953,
+      "step": 5910
+    },
+    {
+      "epoch": 0.3885151763740771,
+      "grad_norm": 0.2839881181716919,
+      "learning_rate": 0.0002,
+      "loss": 0.8789,
+      "step": 5920
+    },
+    {
+      "epoch": 0.38917145200984415,
+      "grad_norm": 0.30544501543045044,
+      "learning_rate": 0.0002,
+      "loss": 0.9585,
+      "step": 5930
+    },
+    {
+      "epoch": 0.38982772764561113,
+      "grad_norm": 0.281946063041687,
+      "learning_rate": 0.0002,
+      "loss": 0.9163,
+      "step": 5940
+    },
+    {
+      "epoch": 0.39048400328137817,
+      "grad_norm": 0.44500306248664856,
+      "learning_rate": 0.0002,
+      "loss": 0.8999,
+      "step": 5950
+    },
+    {
+      "epoch": 0.3911402789171452,
+      "grad_norm": 0.324260950088501,
+      "learning_rate": 0.0002,
+      "loss": 0.8928,
+      "step": 5960
+    },
+    {
+      "epoch": 0.39179655455291224,
+      "grad_norm": 0.31535372138023376,
+      "learning_rate": 0.0002,
+      "loss": 0.8844,
+      "step": 5970
+    },
+    {
+      "epoch": 0.39245283018867927,
+      "grad_norm": 0.31681323051452637,
+      "learning_rate": 0.0002,
+      "loss": 0.9185,
+      "step": 5980
+    },
+    {
+      "epoch": 0.39310910582444625,
+      "grad_norm": 0.32668736577033997,
+      "learning_rate": 0.0002,
+      "loss": 0.9258,
+      "step": 5990
+    },
+    {
+      "epoch": 0.3937653814602133,
+      "grad_norm": 0.34531155228614807,
+      "learning_rate": 0.0002,
+      "loss": 0.9278,
+      "step": 6000
+    },
+    {
+      "epoch": 0.3944216570959803,
+      "grad_norm": 0.32312485575675964,
+      "learning_rate": 0.0002,
+      "loss": 0.8969,
+      "step": 6010
+    },
+    {
+      "epoch": 0.39507793273174735,
+      "grad_norm": 0.31243696808815,
+      "learning_rate": 0.0002,
+      "loss": 0.9501,
+      "step": 6020
+    },
+    {
+      "epoch": 0.39573420836751433,
+      "grad_norm": 0.3558828830718994,
+      "learning_rate": 0.0002,
+      "loss": 0.9174,
+      "step": 6030
+    },
+    {
+      "epoch": 0.39639048400328136,
+      "grad_norm": 0.3894706964492798,
+      "learning_rate": 0.0002,
+      "loss": 0.9486,
+      "step": 6040
+    },
+    {
+      "epoch": 0.3970467596390484,
+      "grad_norm": 0.36445918679237366,
+      "learning_rate": 0.0002,
+      "loss": 0.9549,
+      "step": 6050
+    },
+    {
+      "epoch": 0.39770303527481543,
+      "grad_norm": 0.34586355090141296,
+      "learning_rate": 0.0002,
+      "loss": 0.9268,
+      "step": 6060
+    },
+    {
+      "epoch": 0.39835931091058246,
+      "grad_norm": 0.4428504705429077,
+      "learning_rate": 0.0002,
+      "loss": 0.9256,
+      "step": 6070
+    },
+    {
+      "epoch": 0.39901558654634944,
+      "grad_norm": 0.32784751057624817,
+      "learning_rate": 0.0002,
+      "loss": 0.9706,
+      "step": 6080
+    },
+    {
+      "epoch": 0.3996718621821165,
+      "grad_norm": 0.30480754375457764,
+      "learning_rate": 0.0002,
+      "loss": 0.9107,
+      "step": 6090
+    },
+    {
+      "epoch": 0.4003281378178835,
+      "grad_norm": 0.3382161557674408,
+      "learning_rate": 0.0002,
+      "loss": 0.9431,
+      "step": 6100
+    },
+    {
+      "epoch": 0.40098441345365055,
+      "grad_norm": 0.34486305713653564,
+      "learning_rate": 0.0002,
+      "loss": 0.8669,
+      "step": 6110
+    },
+    {
+      "epoch": 0.4016406890894176,
+      "grad_norm": 0.3488329350948334,
+      "learning_rate": 0.0002,
+      "loss": 0.9174,
+      "step": 6120
+    },
+    {
+      "epoch": 0.40229696472518456,
+      "grad_norm": 0.2977526783943176,
+      "learning_rate": 0.0002,
+      "loss": 0.8907,
+      "step": 6130
+    },
+    {
+      "epoch": 0.4029532403609516,
+      "grad_norm": 0.27584630250930786,
+      "learning_rate": 0.0002,
+      "loss": 0.9088,
+      "step": 6140
+    },
+    {
+      "epoch": 0.4036095159967186,
+      "grad_norm": 0.30325573682785034,
+      "learning_rate": 0.0002,
+      "loss": 0.9062,
+      "step": 6150
+    },
+    {
+      "epoch": 0.40426579163248566,
+      "grad_norm": 0.3318769633769989,
+      "learning_rate": 0.0002,
+      "loss": 0.9626,
+      "step": 6160
+    },
+    {
+      "epoch": 0.40492206726825264,
+      "grad_norm": 0.27929258346557617,
+      "learning_rate": 0.0002,
+      "loss": 0.8849,
+      "step": 6170
+    },
+    {
+      "epoch": 0.4055783429040197,
+      "grad_norm": 0.3127989172935486,
+      "learning_rate": 0.0002,
+      "loss": 0.9118,
+      "step": 6180
+    },
+    {
+      "epoch": 0.4062346185397867,
+      "grad_norm": 0.34189802408218384,
+      "learning_rate": 0.0002,
+      "loss": 0.9659,
+      "step": 6190
+    },
+    {
+      "epoch": 0.40689089417555374,
+      "grad_norm": 0.2901368737220764,
+      "learning_rate": 0.0002,
+      "loss": 0.9041,
+      "step": 6200
+    },
+    {
+      "epoch": 0.4075471698113208,
+      "grad_norm": 0.2957440912723541,
+      "learning_rate": 0.0002,
+      "loss": 0.8863,
+      "step": 6210
+    },
+    {
+      "epoch": 0.40820344544708775,
+      "grad_norm": 0.35854384303092957,
+      "learning_rate": 0.0002,
+      "loss": 0.8872,
+      "step": 6220
+    },
+    {
+      "epoch": 0.4088597210828548,
+      "grad_norm": 0.3516639471054077,
+      "learning_rate": 0.0002,
+      "loss": 0.9913,
+      "step": 6230
+    },
+    {
+      "epoch": 0.4095159967186218,
+      "grad_norm": 0.33773764967918396,
+      "learning_rate": 0.0002,
+      "loss": 0.9526,
+      "step": 6240
+    },
+    {
+      "epoch": 0.41017227235438886,
+      "grad_norm": 0.3359714448451996,
+      "learning_rate": 0.0002,
+      "loss": 0.9116,
+      "step": 6250
+    },
+    {
+      "epoch": 0.4108285479901559,
+      "grad_norm": 0.37056079506874084,
+      "learning_rate": 0.0002,
+      "loss": 0.8789,
+      "step": 6260
+    },
+    {
+      "epoch": 0.41148482362592287,
+      "grad_norm": 0.37948423624038696,
+      "learning_rate": 0.0002,
+      "loss": 0.9262,
+      "step": 6270
+    },
+    {
+      "epoch": 0.4121410992616899,
+      "grad_norm": 0.30971240997314453,
+      "learning_rate": 0.0002,
+      "loss": 0.9333,
+      "step": 6280
+    },
+    {
+      "epoch": 0.41279737489745694,
+      "grad_norm": 0.3357695937156677,
+      "learning_rate": 0.0002,
+      "loss": 0.8636,
+      "step": 6290
+    },
+    {
+      "epoch": 0.41345365053322397,
+      "grad_norm": 0.2873750627040863,
+      "learning_rate": 0.0002,
+      "loss": 0.8857,
+      "step": 6300
+    },
+    {
+      "epoch": 0.41410992616899095,
+      "grad_norm": 0.31098586320877075,
+      "learning_rate": 0.0002,
+      "loss": 0.9401,
+      "step": 6310
+    },
+    {
+      "epoch": 0.414766201804758,
+      "grad_norm": 0.3232232630252838,
+      "learning_rate": 0.0002,
+      "loss": 0.9113,
+      "step": 6320
+    },
+    {
+      "epoch": 0.415422477440525,
+      "grad_norm": 0.3094746172428131,
+      "learning_rate": 0.0002,
+      "loss": 0.9165,
+      "step": 6330
+    },
+    {
+      "epoch": 0.41607875307629205,
+      "grad_norm": 0.3032565414905548,
+      "learning_rate": 0.0002,
+      "loss": 0.92,
+      "step": 6340
+    },
+    {
+      "epoch": 0.4167350287120591,
+      "grad_norm": 0.30271366238594055,
+      "learning_rate": 0.0002,
+      "loss": 0.9254,
+      "step": 6350
+    },
+    {
+      "epoch": 0.41739130434782606,
+      "grad_norm": 0.35404741764068604,
+      "learning_rate": 0.0002,
+      "loss": 0.9766,
+      "step": 6360
+    },
+    {
+      "epoch": 0.4180475799835931,
+      "grad_norm": 0.33428773283958435,
+      "learning_rate": 0.0002,
+      "loss": 0.889,
+      "step": 6370
+    },
+    {
+      "epoch": 0.41870385561936013,
+      "grad_norm": 0.3523367941379547,
+      "learning_rate": 0.0002,
+      "loss": 0.9284,
+      "step": 6380
+    },
+    {
+      "epoch": 0.41936013125512717,
+      "grad_norm": 0.28411954641342163,
+      "learning_rate": 0.0002,
+      "loss": 0.8831,
+      "step": 6390
+    },
+    {
+      "epoch": 0.4200164068908942,
+      "grad_norm": 0.3410111963748932,
+      "learning_rate": 0.0002,
+      "loss": 0.9429,
+      "step": 6400
+    },
+    {
+      "epoch": 0.4206726825266612,
+      "grad_norm": 0.33837398886680603,
+      "learning_rate": 0.0002,
+      "loss": 0.8956,
+      "step": 6410
+    },
+    {
+      "epoch": 0.4213289581624282,
+      "grad_norm": 0.38739341497421265,
+      "learning_rate": 0.0002,
+      "loss": 0.9169,
+      "step": 6420
+    },
+    {
+      "epoch": 0.42198523379819525,
+      "grad_norm": 0.28810667991638184,
+      "learning_rate": 0.0002,
+      "loss": 0.8943,
+      "step": 6430
+    },
+    {
+      "epoch": 0.4226415094339623,
+      "grad_norm": 0.3242197036743164,
+      "learning_rate": 0.0002,
+      "loss": 0.9505,
+      "step": 6440
+    },
+    {
+      "epoch": 0.42329778506972926,
+      "grad_norm": 0.39538851380348206,
+      "learning_rate": 0.0002,
+      "loss": 0.9409,
+      "step": 6450
+    },
+    {
+      "epoch": 0.4239540607054963,
+      "grad_norm": 0.36792051792144775,
+      "learning_rate": 0.0002,
+      "loss": 0.9152,
+      "step": 6460
+    },
+    {
+      "epoch": 0.42461033634126333,
+      "grad_norm": 0.31599462032318115,
+      "learning_rate": 0.0002,
+      "loss": 0.9203,
+      "step": 6470
+    },
+    {
+      "epoch": 0.42526661197703036,
+      "grad_norm": 0.3358552157878876,
+      "learning_rate": 0.0002,
+      "loss": 0.926,
+      "step": 6480
+    },
+    {
+      "epoch": 0.4259228876127974,
+      "grad_norm": 0.40409335494041443,
+      "learning_rate": 0.0002,
+      "loss": 0.9337,
+      "step": 6490
+    },
+    {
+      "epoch": 0.4265791632485644,
+      "grad_norm": 0.3785623610019684,
+      "learning_rate": 0.0002,
+      "loss": 0.901,
+      "step": 6500
+    },
+    {
+      "epoch": 0.4272354388843314,
+      "grad_norm": 0.43033692240715027,
+      "learning_rate": 0.0002,
+      "loss": 0.9281,
+      "step": 6510
+    },
+    {
+      "epoch": 0.42789171452009844,
+      "grad_norm": 0.3074797987937927,
+      "learning_rate": 0.0002,
+      "loss": 0.9332,
+      "step": 6520
+    },
+    {
+      "epoch": 0.4285479901558655,
+      "grad_norm": 0.3498784005641937,
+      "learning_rate": 0.0002,
+      "loss": 0.9161,
+      "step": 6530
+    },
+    {
+      "epoch": 0.4292042657916325,
+      "grad_norm": 0.40507981181144714,
+      "learning_rate": 0.0002,
+      "loss": 0.9829,
+      "step": 6540
+    },
+    {
+      "epoch": 0.4298605414273995,
+      "grad_norm": 0.35614013671875,
+      "learning_rate": 0.0002,
+      "loss": 0.9269,
+      "step": 6550
+    },
+    {
+      "epoch": 0.4305168170631665,
+      "grad_norm": 0.451616495847702,
+      "learning_rate": 0.0002,
+      "loss": 0.9033,
+      "step": 6560
+    },
+    {
+      "epoch": 0.43117309269893356,
+      "grad_norm": 0.3707262873649597,
+      "learning_rate": 0.0002,
+      "loss": 0.9723,
+      "step": 6570
+    },
+    {
+      "epoch": 0.4318293683347006,
+      "grad_norm": 0.3271283209323883,
+      "learning_rate": 0.0002,
+      "loss": 0.9512,
+      "step": 6580
+    },
+    {
+      "epoch": 0.43248564397046757,
+      "grad_norm": 0.38105711340904236,
+      "learning_rate": 0.0002,
+      "loss": 0.8529,
+      "step": 6590
+    },
+    {
+      "epoch": 0.4331419196062346,
+      "grad_norm": 0.345217227935791,
+      "learning_rate": 0.0002,
+      "loss": 0.9208,
+      "step": 6600
+    },
+    {
+      "epoch": 0.43379819524200164,
+      "grad_norm": 0.3507174551486969,
+      "learning_rate": 0.0002,
+      "loss": 0.8958,
+      "step": 6610
+    },
+    {
+      "epoch": 0.4344544708777687,
+      "grad_norm": 0.318654328584671,
+      "learning_rate": 0.0002,
+      "loss": 0.9402,
+      "step": 6620
+    },
+    {
+      "epoch": 0.4351107465135357,
+      "grad_norm": 0.3628910183906555,
+      "learning_rate": 0.0002,
+      "loss": 0.9068,
+      "step": 6630
+    },
+    {
+      "epoch": 0.4357670221493027,
+      "grad_norm": 0.4454437494277954,
+      "learning_rate": 0.0002,
+      "loss": 0.9122,
+      "step": 6640
+    },
+    {
+      "epoch": 0.4364232977850697,
+      "grad_norm": 0.36227893829345703,
+      "learning_rate": 0.0002,
+      "loss": 1.0092,
+      "step": 6650
+    },
+    {
+      "epoch": 0.43707957342083675,
+      "grad_norm": 0.33995604515075684,
+      "learning_rate": 0.0002,
+      "loss": 0.9289,
+      "step": 6660
+    },
+    {
+      "epoch": 0.4377358490566038,
+      "grad_norm": 0.3784409165382385,
+      "learning_rate": 0.0002,
+      "loss": 0.9032,
+      "step": 6670
+    },
+    {
+      "epoch": 0.4383921246923708,
+      "grad_norm": 0.39470377564430237,
+      "learning_rate": 0.0002,
+      "loss": 0.9399,
+      "step": 6680
+    },
+    {
+      "epoch": 0.4390484003281378,
+      "grad_norm": 0.34198591113090515,
+      "learning_rate": 0.0002,
+      "loss": 0.8795,
+      "step": 6690
+    },
+    {
+      "epoch": 0.43970467596390483,
+      "grad_norm": 0.35575005412101746,
+      "learning_rate": 0.0002,
+      "loss": 0.9665,
+      "step": 6700
+    },
+    {
+      "epoch": 0.44036095159967187,
+      "grad_norm": 0.40915772318840027,
+      "learning_rate": 0.0002,
+      "loss": 0.9133,
+      "step": 6710
+    },
+    {
+      "epoch": 0.4410172272354389,
+      "grad_norm": 0.3526946008205414,
+      "learning_rate": 0.0002,
+      "loss": 0.9112,
+      "step": 6720
+    },
+    {
+      "epoch": 0.4416735028712059,
+      "grad_norm": 0.30411285161972046,
+      "learning_rate": 0.0002,
+      "loss": 0.8886,
+      "step": 6730
+    },
+    {
+      "epoch": 0.4423297785069729,
+      "grad_norm": 0.2935996949672699,
+      "learning_rate": 0.0002,
+      "loss": 0.8969,
+      "step": 6740
+    },
+    {
+      "epoch": 0.44298605414273995,
+      "grad_norm": 0.3806987702846527,
+      "learning_rate": 0.0002,
+      "loss": 0.8864,
+      "step": 6750
+    },
+    {
+      "epoch": 0.443642329778507,
+      "grad_norm": 0.3329904079437256,
+      "learning_rate": 0.0002,
+      "loss": 0.9267,
+      "step": 6760
+    },
+    {
+      "epoch": 0.444298605414274,
+      "grad_norm": 0.33832767605781555,
+      "learning_rate": 0.0002,
+      "loss": 0.9223,
+      "step": 6770
+    },
+    {
+      "epoch": 0.444954881050041,
+      "grad_norm": 0.3387627899646759,
+      "learning_rate": 0.0002,
+      "loss": 0.9127,
+      "step": 6780
+    },
+    {
+      "epoch": 0.44561115668580803,
+      "grad_norm": 0.32709822058677673,
+      "learning_rate": 0.0002,
+      "loss": 0.8853,
+      "step": 6790
+    },
+    {
+      "epoch": 0.44626743232157506,
+      "grad_norm": 0.2914820909500122,
+      "learning_rate": 0.0002,
+      "loss": 0.9059,
+      "step": 6800
+    },
+    {
+      "epoch": 0.4469237079573421,
+      "grad_norm": 0.3288695216178894,
+      "learning_rate": 0.0002,
+      "loss": 0.9209,
+      "step": 6810
+    },
+    {
+      "epoch": 0.44757998359310913,
+      "grad_norm": 0.5003459453582764,
+      "learning_rate": 0.0002,
+      "loss": 0.9208,
+      "step": 6820
+    },
+    {
+      "epoch": 0.4482362592288761,
+      "grad_norm": 0.359402060508728,
+      "learning_rate": 0.0002,
+      "loss": 0.9797,
+      "step": 6830
+    },
+    {
+      "epoch": 0.44889253486464314,
+      "grad_norm": 0.35508984327316284,
+      "learning_rate": 0.0002,
+      "loss": 0.856,
+      "step": 6840
+    },
+    {
+      "epoch": 0.4495488105004102,
+      "grad_norm": 0.33327680826187134,
+      "learning_rate": 0.0002,
+      "loss": 0.9513,
+      "step": 6850
+    },
+    {
+      "epoch": 0.4502050861361772,
+      "grad_norm": 0.4084452688694,
+      "learning_rate": 0.0002,
+      "loss": 0.965,
+      "step": 6860
+    },
+    {
+      "epoch": 0.4508613617719442,
+      "grad_norm": 0.30232545733451843,
+      "learning_rate": 0.0002,
+      "loss": 0.899,
+      "step": 6870
+    },
+    {
+      "epoch": 0.4515176374077112,
+      "grad_norm": 0.3244702219963074,
+      "learning_rate": 0.0002,
+      "loss": 0.9294,
+      "step": 6880
+    },
+    {
+      "epoch": 0.45217391304347826,
+      "grad_norm": 0.32148104906082153,
+      "learning_rate": 0.0002,
+      "loss": 0.9312,
+      "step": 6890
+    },
+    {
+      "epoch": 0.4528301886792453,
+      "grad_norm": 0.3831937313079834,
+      "learning_rate": 0.0002,
+      "loss": 1.007,
+      "step": 6900
+    },
+    {
+      "epoch": 0.4534864643150123,
+      "grad_norm": 0.33381497859954834,
+      "learning_rate": 0.0002,
+      "loss": 0.9111,
+      "step": 6910
+    },
+    {
+      "epoch": 0.4541427399507793,
+      "grad_norm": 0.4074220359325409,
+      "learning_rate": 0.0002,
+      "loss": 0.8955,
+      "step": 6920
+    },
+    {
+      "epoch": 0.45479901558654634,
+      "grad_norm": 0.35602903366088867,
+      "learning_rate": 0.0002,
+      "loss": 0.9326,
+      "step": 6930
+    },
+    {
+      "epoch": 0.4554552912223134,
+      "grad_norm": 0.3411916196346283,
+      "learning_rate": 0.0002,
+      "loss": 0.9146,
+      "step": 6940
+    },
+    {
+      "epoch": 0.4561115668580804,
+      "grad_norm": 0.38504868745803833,
+      "learning_rate": 0.0002,
+      "loss": 0.9382,
+      "step": 6950
+    },
+    {
+      "epoch": 0.45676784249384744,
+      "grad_norm": 0.5072926878929138,
+      "learning_rate": 0.0002,
+      "loss": 0.9105,
+      "step": 6960
+    },
+    {
+      "epoch": 0.4574241181296144,
+      "grad_norm": 0.38257330656051636,
+      "learning_rate": 0.0002,
+      "loss": 0.9383,
+      "step": 6970
+    },
+    {
+      "epoch": 0.45808039376538146,
+      "grad_norm": 0.40472176671028137,
+      "learning_rate": 0.0002,
+      "loss": 0.8864,
+      "step": 6980
+    },
+    {
+      "epoch": 0.4587366694011485,
+      "grad_norm": 0.34501466155052185,
+      "learning_rate": 0.0002,
+      "loss": 0.949,
+      "step": 6990
+    },
+    {
+      "epoch": 0.4593929450369155,
+      "grad_norm": 0.32561028003692627,
+      "learning_rate": 0.0002,
+      "loss": 0.9717,
+      "step": 7000
+    },
+    {
+      "epoch": 0.4600492206726825,
+      "grad_norm": 0.3388477861881256,
+      "learning_rate": 0.0002,
+      "loss": 0.9655,
+      "step": 7010
+    },
+    {
+      "epoch": 0.46070549630844954,
+      "grad_norm": 0.38045769929885864,
+      "learning_rate": 0.0002,
+      "loss": 0.9039,
+      "step": 7020
+    },
+    {
+      "epoch": 0.46136177194421657,
+      "grad_norm": 0.32500898838043213,
+      "learning_rate": 0.0002,
+      "loss": 0.8879,
+      "step": 7030
+    },
+    {
+      "epoch": 0.4620180475799836,
+      "grad_norm": 0.3758746385574341,
+      "learning_rate": 0.0002,
+      "loss": 0.9872,
+      "step": 7040
+    },
+    {
+      "epoch": 0.46267432321575064,
+      "grad_norm": 0.39033588767051697,
+      "learning_rate": 0.0002,
+      "loss": 0.9043,
+      "step": 7050
+    },
+    {
+      "epoch": 0.4633305988515176,
+      "grad_norm": 0.35141658782958984,
+      "learning_rate": 0.0002,
+      "loss": 0.891,
+      "step": 7060
+    },
+    {
+      "epoch": 0.46398687448728465,
+      "grad_norm": 0.3715546727180481,
+      "learning_rate": 0.0002,
+      "loss": 0.9196,
+      "step": 7070
+    },
+    {
+      "epoch": 0.4646431501230517,
+      "grad_norm": 0.3062947988510132,
+      "learning_rate": 0.0002,
+      "loss": 0.9051,
+      "step": 7080
+    },
+    {
+      "epoch": 0.4652994257588187,
+      "grad_norm": 0.43001696467399597,
+      "learning_rate": 0.0002,
+      "loss": 0.9395,
+      "step": 7090
+    },
+    {
+      "epoch": 0.46595570139458575,
+      "grad_norm": 0.34747597575187683,
+      "learning_rate": 0.0002,
+      "loss": 0.9363,
+      "step": 7100
+    },
+    {
+      "epoch": 0.46661197703035273,
+      "grad_norm": 0.33514827489852905,
+      "learning_rate": 0.0002,
+      "loss": 0.9052,
+      "step": 7110
+    },
+    {
+      "epoch": 0.46726825266611977,
+      "grad_norm": 0.3500545918941498,
+      "learning_rate": 0.0002,
+      "loss": 0.8673,
+      "step": 7120
+    },
+    {
+      "epoch": 0.4679245283018868,
+      "grad_norm": 0.3959280252456665,
+      "learning_rate": 0.0002,
+      "loss": 0.9373,
+      "step": 7130
+    },
+    {
+      "epoch": 0.46858080393765383,
+      "grad_norm": 0.3119291067123413,
+      "learning_rate": 0.0002,
+      "loss": 0.868,
+      "step": 7140
+    },
+    {
+      "epoch": 0.4692370795734208,
+      "grad_norm": 0.36544349789619446,
+      "learning_rate": 0.0002,
+      "loss": 0.923,
+      "step": 7150
+    },
+    {
+      "epoch": 0.46989335520918785,
+      "grad_norm": 0.3375662565231323,
+      "learning_rate": 0.0002,
+      "loss": 0.9281,
+      "step": 7160
+    },
+    {
+      "epoch": 0.4705496308449549,
+      "grad_norm": 0.48938584327697754,
+      "learning_rate": 0.0002,
+      "loss": 0.8948,
+      "step": 7170
+    },
+    {
+      "epoch": 0.4712059064807219,
+      "grad_norm": 0.36706942319869995,
+      "learning_rate": 0.0002,
+      "loss": 0.9016,
+      "step": 7180
+    },
+    {
+      "epoch": 0.47186218211648895,
+      "grad_norm": 0.43472692370414734,
+      "learning_rate": 0.0002,
+      "loss": 0.906,
+      "step": 7190
+    },
+    {
+      "epoch": 0.4725184577522559,
+      "grad_norm": 0.3433953821659088,
+      "learning_rate": 0.0002,
+      "loss": 0.9114,
+      "step": 7200
+    },
+    {
+      "epoch": 0.47317473338802296,
+      "grad_norm": 0.3224952518939972,
+      "learning_rate": 0.0002,
+      "loss": 0.899,
+      "step": 7210
+    },
+    {
+      "epoch": 0.47383100902379,
+      "grad_norm": 0.34640923142433167,
+      "learning_rate": 0.0002,
+      "loss": 0.9039,
+      "step": 7220
+    },
+    {
+      "epoch": 0.47448728465955703,
+      "grad_norm": 0.36221131682395935,
+      "learning_rate": 0.0002,
+      "loss": 0.9372,
+      "step": 7230
+    },
+    {
+      "epoch": 0.47514356029532406,
+      "grad_norm": 0.3695160150527954,
+      "learning_rate": 0.0002,
+      "loss": 0.9328,
+      "step": 7240
+    },
+    {
+      "epoch": 0.47579983593109104,
+      "grad_norm": 0.3915520906448364,
+      "learning_rate": 0.0002,
+      "loss": 0.8726,
+      "step": 7250
+    },
+    {
+      "epoch": 0.4764561115668581,
+      "grad_norm": 0.34111252427101135,
+      "learning_rate": 0.0002,
+      "loss": 0.9305,
+      "step": 7260
+    },
+    {
+      "epoch": 0.4771123872026251,
+      "grad_norm": 0.3156047761440277,
+      "learning_rate": 0.0002,
+      "loss": 0.8853,
+      "step": 7270
+    },
+    {
+      "epoch": 0.47776866283839214,
+      "grad_norm": 0.349992573261261,
+      "learning_rate": 0.0002,
+      "loss": 0.9442,
+      "step": 7280
+    },
+    {
+      "epoch": 0.4784249384741591,
+      "grad_norm": 0.3627530634403229,
+      "learning_rate": 0.0002,
+      "loss": 0.8841,
+      "step": 7290
+    },
+    {
+      "epoch": 0.47908121410992616,
+      "grad_norm": 0.39342308044433594,
+      "learning_rate": 0.0002,
+      "loss": 0.937,
+      "step": 7300
+    },
+    {
+      "epoch": 0.4797374897456932,
+      "grad_norm": 0.321122944355011,
+      "learning_rate": 0.0002,
+      "loss": 0.951,
+      "step": 7310
+    },
+    {
+      "epoch": 0.4803937653814602,
+      "grad_norm": 0.37064728140830994,
+      "learning_rate": 0.0002,
+      "loss": 0.9085,
+      "step": 7320
+    },
+    {
+      "epoch": 0.48105004101722726,
+      "grad_norm": 0.348779559135437,
+      "learning_rate": 0.0002,
+      "loss": 0.9479,
+      "step": 7330
+    },
+    {
+      "epoch": 0.48170631665299424,
+      "grad_norm": 0.3484368324279785,
+      "learning_rate": 0.0002,
+      "loss": 0.8651,
+      "step": 7340
+    },
+    {
+      "epoch": 0.48236259228876127,
+      "grad_norm": 0.3343070149421692,
+      "learning_rate": 0.0002,
+      "loss": 0.9096,
+      "step": 7350
+    },
+    {
+      "epoch": 0.4830188679245283,
+      "grad_norm": 0.3310532867908478,
+      "learning_rate": 0.0002,
+      "loss": 0.9148,
+      "step": 7360
+    },
+    {
+      "epoch": 0.48367514356029534,
+      "grad_norm": 0.37587985396385193,
+      "learning_rate": 0.0002,
+      "loss": 0.8652,
+      "step": 7370
+    },
+    {
+      "epoch": 0.4843314191960624,
+      "grad_norm": 0.37325888872146606,
+      "learning_rate": 0.0002,
+      "loss": 0.9302,
+      "step": 7380
+    },
+    {
+      "epoch": 0.48498769483182935,
+      "grad_norm": 0.3869538903236389,
+      "learning_rate": 0.0002,
+      "loss": 0.9382,
+      "step": 7390
+    },
+    {
+      "epoch": 0.4856439704675964,
+      "grad_norm": 0.34064534306526184,
+      "learning_rate": 0.0002,
+      "loss": 0.8967,
+      "step": 7400
+    },
+    {
+      "epoch": 0.4863002461033634,
+      "grad_norm": 0.35124653577804565,
+      "learning_rate": 0.0002,
+      "loss": 0.8884,
+      "step": 7410
+    },
+    {
+      "epoch": 0.48695652173913045,
+      "grad_norm": 0.38339361548423767,
+      "learning_rate": 0.0002,
+      "loss": 0.9178,
+      "step": 7420
+    },
+    {
+      "epoch": 0.48761279737489743,
+      "grad_norm": 0.32823026180267334,
+      "learning_rate": 0.0002,
+      "loss": 0.8984,
+      "step": 7430
+    },
+    {
+      "epoch": 0.48826907301066447,
+      "grad_norm": 0.32759153842926025,
+      "learning_rate": 0.0002,
+      "loss": 0.9348,
+      "step": 7440
+    },
+    {
+      "epoch": 0.4889253486464315,
+      "grad_norm": 0.3855777680873871,
+      "learning_rate": 0.0002,
+      "loss": 0.9023,
+      "step": 7450
+    },
+    {
+      "epoch": 0.48958162428219854,
+      "grad_norm": 0.34890854358673096,
+      "learning_rate": 0.0002,
+      "loss": 0.9201,
+      "step": 7460
+    },
+    {
+      "epoch": 0.49023789991796557,
+      "grad_norm": 0.3580712378025055,
+      "learning_rate": 0.0002,
+      "loss": 0.8947,
+      "step": 7470
+    },
+    {
+      "epoch": 0.49089417555373255,
+      "grad_norm": 0.34897708892822266,
+      "learning_rate": 0.0002,
+      "loss": 0.8413,
+      "step": 7480
+    },
+    {
+      "epoch": 0.4915504511894996,
+      "grad_norm": 0.32964155077934265,
+      "learning_rate": 0.0002,
+      "loss": 0.8508,
+      "step": 7490
+    },
+    {
+      "epoch": 0.4922067268252666,
+      "grad_norm": 0.3805377185344696,
+      "learning_rate": 0.0002,
+      "loss": 0.9298,
+      "step": 7500
+    },
+    {
+      "epoch": 0.49286300246103365,
+      "grad_norm": 0.3312858045101166,
+      "learning_rate": 0.0002,
+      "loss": 0.9354,
+      "step": 7510
+    },
+    {
+      "epoch": 0.49351927809680063,
+      "grad_norm": 0.32818418741226196,
+      "learning_rate": 0.0002,
+      "loss": 0.8966,
+      "step": 7520
+    },
+    {
+      "epoch": 0.49417555373256766,
+      "grad_norm": 0.3721756339073181,
+      "learning_rate": 0.0002,
+      "loss": 0.9036,
+      "step": 7530
+    },
+    {
+      "epoch": 0.4948318293683347,
+      "grad_norm": 0.4272252321243286,
+      "learning_rate": 0.0002,
+      "loss": 0.9483,
+      "step": 7540
+    },
+    {
+      "epoch": 0.49548810500410173,
+      "grad_norm": 0.3034988343715668,
+      "learning_rate": 0.0002,
+      "loss": 0.9159,
+      "step": 7550
+    },
+    {
+      "epoch": 0.49614438063986877,
+      "grad_norm": 0.34006139636039734,
+      "learning_rate": 0.0002,
+      "loss": 0.8996,
+      "step": 7560
+    },
+    {
+      "epoch": 0.49680065627563574,
+      "grad_norm": 0.33074137568473816,
+      "learning_rate": 0.0002,
+      "loss": 0.9122,
+      "step": 7570
+    },
+    {
+      "epoch": 0.4974569319114028,
+      "grad_norm": 0.3747742772102356,
+      "learning_rate": 0.0002,
+      "loss": 0.877,
+      "step": 7580
+    },
+    {
+      "epoch": 0.4981132075471698,
+      "grad_norm": 0.29682815074920654,
+      "learning_rate": 0.0002,
+      "loss": 0.8738,
+      "step": 7590
+    },
+    {
+      "epoch": 0.49876948318293685,
+      "grad_norm": 0.36150112748146057,
+      "learning_rate": 0.0002,
+      "loss": 0.9669,
+      "step": 7600
+    },
+    {
+      "epoch": 0.4994257588187039,
+      "grad_norm": 0.3893260955810547,
+      "learning_rate": 0.0002,
+      "loss": 0.8936,
+      "step": 7610
+    },
+    {
+      "epoch": 0.5000820344544709,
+      "grad_norm": 0.358791708946228,
+      "learning_rate": 0.0002,
+      "loss": 0.9273,
+      "step": 7620
+    },
+    {
+      "epoch": 0.500738310090238,
+      "grad_norm": 0.3799469470977783,
+      "learning_rate": 0.0002,
+      "loss": 0.9053,
+      "step": 7630
+    },
+    {
+      "epoch": 0.5013945857260049,
+      "grad_norm": 0.35908520221710205,
+      "learning_rate": 0.0002,
+      "loss": 0.9655,
+      "step": 7640
+    },
+    {
+      "epoch": 0.5020508613617719,
+      "grad_norm": 0.33538103103637695,
+      "learning_rate": 0.0002,
+      "loss": 0.8993,
+      "step": 7650
+    },
+    {
+      "epoch": 0.502707136997539,
+      "grad_norm": 0.2990545928478241,
+      "learning_rate": 0.0002,
+      "loss": 0.8812,
+      "step": 7660
+    },
+    {
+      "epoch": 0.503363412633306,
+      "grad_norm": 0.3342699408531189,
+      "learning_rate": 0.0002,
+      "loss": 0.8965,
+      "step": 7670
+    },
+    {
+      "epoch": 0.5040196882690731,
+      "grad_norm": 0.3656594455242157,
+      "learning_rate": 0.0002,
+      "loss": 0.9476,
+      "step": 7680
+    },
+    {
+      "epoch": 0.50467596390484,
+      "grad_norm": 0.315176784992218,
+      "learning_rate": 0.0002,
+      "loss": 0.9225,
+      "step": 7690
+    },
+    {
+      "epoch": 0.505332239540607,
+      "grad_norm": 0.3070623576641083,
+      "learning_rate": 0.0002,
+      "loss": 0.8998,
+      "step": 7700
+    },
+    {
+      "epoch": 0.5059885151763741,
+      "grad_norm": 0.3741548955440521,
+      "learning_rate": 0.0002,
+      "loss": 0.8565,
+      "step": 7710
+    },
+    {
+      "epoch": 0.5066447908121411,
+      "grad_norm": 0.3047017753124237,
+      "learning_rate": 0.0002,
+      "loss": 0.8882,
+      "step": 7720
+    },
+    {
+      "epoch": 0.5073010664479081,
+      "grad_norm": 0.5524629354476929,
+      "learning_rate": 0.0002,
+      "loss": 0.9002,
+      "step": 7730
+    },
+    {
+      "epoch": 0.5079573420836752,
+      "grad_norm": 0.39234456419944763,
+      "learning_rate": 0.0002,
+      "loss": 0.901,
+      "step": 7740
+    },
+    {
+      "epoch": 0.5086136177194421,
+      "grad_norm": 0.3306262791156769,
+      "learning_rate": 0.0002,
+      "loss": 0.8948,
+      "step": 7750
+    },
+    {
+      "epoch": 0.5092698933552092,
+      "grad_norm": 0.3181036710739136,
+      "learning_rate": 0.0002,
+      "loss": 0.8627,
+      "step": 7760
+    },
+    {
+      "epoch": 0.5099261689909762,
+      "grad_norm": 0.3531821668148041,
+      "learning_rate": 0.0002,
+      "loss": 0.9068,
+      "step": 7770
+    },
+    {
+      "epoch": 0.5105824446267432,
+      "grad_norm": 0.39781567454338074,
+      "learning_rate": 0.0002,
+      "loss": 0.9452,
+      "step": 7780
+    },
+    {
+      "epoch": 0.5112387202625103,
+      "grad_norm": 0.32413530349731445,
+      "learning_rate": 0.0002,
+      "loss": 0.8829,
+      "step": 7790
+    },
+    {
+      "epoch": 0.5118949958982772,
+      "grad_norm": 0.33857491612434387,
+      "learning_rate": 0.0002,
+      "loss": 0.9162,
+      "step": 7800
+    },
+    {
+      "epoch": 0.5125512715340443,
+      "grad_norm": 0.33426207304000854,
+      "learning_rate": 0.0002,
+      "loss": 0.9183,
+      "step": 7810
+    },
+    {
+      "epoch": 0.5132075471698113,
+      "grad_norm": 0.31170961260795593,
+      "learning_rate": 0.0002,
+      "loss": 0.8487,
+      "step": 7820
+    },
+    {
+      "epoch": 0.5138638228055783,
+      "grad_norm": 0.3401695787906647,
+      "learning_rate": 0.0002,
+      "loss": 0.9425,
+      "step": 7830
+    },
+    {
+      "epoch": 0.5145200984413454,
+      "grad_norm": 0.4325079619884491,
+      "learning_rate": 0.0002,
+      "loss": 0.8852,
+      "step": 7840
+    },
+    {
+      "epoch": 0.5151763740771124,
+      "grad_norm": 0.33266621828079224,
+      "learning_rate": 0.0002,
+      "loss": 0.9165,
+      "step": 7850
+    },
+    {
+      "epoch": 0.5158326497128795,
+      "grad_norm": 0.353476345539093,
+      "learning_rate": 0.0002,
+      "loss": 0.9178,
+      "step": 7860
+    },
+    {
+      "epoch": 0.5164889253486464,
+      "grad_norm": 0.34789302945137024,
+      "learning_rate": 0.0002,
+      "loss": 0.9734,
+      "step": 7870
+    },
+    {
+      "epoch": 0.5171452009844134,
+      "grad_norm": 0.3378570079803467,
+      "learning_rate": 0.0002,
+      "loss": 0.8674,
+      "step": 7880
+    },
+    {
+      "epoch": 0.5178014766201805,
+      "grad_norm": 0.340122789144516,
+      "learning_rate": 0.0002,
+      "loss": 0.9147,
+      "step": 7890
+    },
+    {
+      "epoch": 0.5184577522559475,
+      "grad_norm": 0.39235031604766846,
+      "learning_rate": 0.0002,
+      "loss": 0.9627,
+      "step": 7900
+    },
+    {
+      "epoch": 0.5191140278917146,
+      "grad_norm": 0.31752005219459534,
+      "learning_rate": 0.0002,
+      "loss": 0.9003,
+      "step": 7910
+    },
+    {
+      "epoch": 0.5197703035274815,
+      "grad_norm": 0.3948894739151001,
+      "learning_rate": 0.0002,
+      "loss": 0.9353,
+      "step": 7920
+    },
+    {
+      "epoch": 0.5204265791632485,
+      "grad_norm": 0.29803305864334106,
+      "learning_rate": 0.0002,
+      "loss": 0.8697,
+      "step": 7930
+    },
+    {
+      "epoch": 0.5210828547990156,
+      "grad_norm": 0.34285855293273926,
+      "learning_rate": 0.0002,
+      "loss": 0.9912,
+      "step": 7940
+    },
+    {
+      "epoch": 0.5217391304347826,
+      "grad_norm": 0.40978604555130005,
+      "learning_rate": 0.0002,
+      "loss": 0.9241,
+      "step": 7950
+    },
+    {
+      "epoch": 0.5223954060705497,
+      "grad_norm": 0.35473906993865967,
+      "learning_rate": 0.0002,
+      "loss": 0.9609,
+      "step": 7960
+    },
+    {
+      "epoch": 0.5230516817063167,
+      "grad_norm": 0.3896619379520416,
+      "learning_rate": 0.0002,
+      "loss": 0.925,
+      "step": 7970
+    },
+    {
+      "epoch": 0.5237079573420836,
+      "grad_norm": 0.3622605502605438,
+      "learning_rate": 0.0002,
+      "loss": 0.8863,
+      "step": 7980
+    },
+    {
+      "epoch": 0.5243642329778507,
+      "grad_norm": 0.4097590744495392,
+      "learning_rate": 0.0002,
+      "loss": 0.939,
+      "step": 7990
+    },
+    {
+      "epoch": 0.5250205086136177,
+      "grad_norm": 0.31397542357444763,
+      "learning_rate": 0.0002,
+      "loss": 0.8895,
+      "step": 8000
+    },
+    {
+      "epoch": 0.5256767842493847,
+      "grad_norm": 0.45559775829315186,
+      "learning_rate": 0.0002,
+      "loss": 0.9185,
+      "step": 8010
+    },
+    {
+      "epoch": 0.5263330598851518,
+      "grad_norm": 0.3471437692642212,
+      "learning_rate": 0.0002,
+      "loss": 0.9318,
+      "step": 8020
+    },
+    {
+      "epoch": 0.5269893355209188,
+      "grad_norm": 0.3105354607105255,
+      "learning_rate": 0.0002,
+      "loss": 0.8684,
+      "step": 8030
+    },
+    {
+      "epoch": 0.5276456111566858,
+      "grad_norm": 0.42847758531570435,
+      "learning_rate": 0.0002,
+      "loss": 0.8633,
+      "step": 8040
+    },
+    {
+      "epoch": 0.5283018867924528,
+      "grad_norm": 0.732471227645874,
+      "learning_rate": 0.0002,
+      "loss": 0.924,
+      "step": 8050
+    },
+    {
+      "epoch": 0.5289581624282198,
+      "grad_norm": 0.3242183327674866,
+      "learning_rate": 0.0002,
+      "loss": 0.8332,
+      "step": 8060
+    },
+    {
+      "epoch": 0.5296144380639869,
+      "grad_norm": 0.36898288130760193,
+      "learning_rate": 0.0002,
+      "loss": 0.9531,
+      "step": 8070
+    },
+    {
+      "epoch": 0.5302707136997539,
+      "grad_norm": 0.3850383162498474,
+      "learning_rate": 0.0002,
+      "loss": 0.933,
+      "step": 8080
+    },
+    {
+      "epoch": 0.530926989335521,
+      "grad_norm": 0.3375824987888336,
+      "learning_rate": 0.0002,
+      "loss": 0.9184,
+      "step": 8090
+    },
+    {
+      "epoch": 0.5315832649712879,
+      "grad_norm": 0.4321442246437073,
+      "learning_rate": 0.0002,
+      "loss": 0.9284,
+      "step": 8100
+    },
+    {
+      "epoch": 0.5322395406070549,
+      "grad_norm": 0.3836155831813812,
+      "learning_rate": 0.0002,
+      "loss": 0.9389,
+      "step": 8110
+    },
+    {
+      "epoch": 0.532895816242822,
+      "grad_norm": 0.36536362767219543,
+      "learning_rate": 0.0002,
+      "loss": 0.9117,
+      "step": 8120
+    },
+    {
+      "epoch": 0.533552091878589,
+      "grad_norm": 0.4364359676837921,
+      "learning_rate": 0.0002,
+      "loss": 0.8724,
+      "step": 8130
+    },
+    {
+      "epoch": 0.5342083675143561,
+      "grad_norm": 0.413506418466568,
+      "learning_rate": 0.0002,
+      "loss": 0.8866,
+      "step": 8140
+    },
+    {
+      "epoch": 0.534864643150123,
+      "grad_norm": 0.33120569586753845,
+      "learning_rate": 0.0002,
+      "loss": 0.8985,
+      "step": 8150
+    },
+    {
+      "epoch": 0.53552091878589,
+      "grad_norm": 0.3408608138561249,
+      "learning_rate": 0.0002,
+      "loss": 0.9553,
+      "step": 8160
+    },
+    {
+      "epoch": 0.5361771944216571,
+      "grad_norm": 0.30265191197395325,
+      "learning_rate": 0.0002,
+      "loss": 0.888,
+      "step": 8170
+    },
+    {
+      "epoch": 0.5368334700574241,
+      "grad_norm": 0.40208759903907776,
+      "learning_rate": 0.0002,
+      "loss": 0.8799,
+      "step": 8180
+    },
+    {
+      "epoch": 0.5374897456931912,
+      "grad_norm": 0.38238683342933655,
+      "learning_rate": 0.0002,
+      "loss": 0.9343,
+      "step": 8190
+    },
+    {
+      "epoch": 0.5381460213289582,
+      "grad_norm": 0.3405042290687561,
+      "learning_rate": 0.0002,
+      "loss": 0.9051,
+      "step": 8200
+    },
+    {
+      "epoch": 0.5388022969647251,
+      "grad_norm": 0.3806214928627014,
+      "learning_rate": 0.0002,
+      "loss": 0.9355,
+      "step": 8210
+    },
+    {
+      "epoch": 0.5394585726004922,
+      "grad_norm": 0.3639737069606781,
+      "learning_rate": 0.0002,
+      "loss": 0.9148,
+      "step": 8220
+    },
+    {
+      "epoch": 0.5401148482362592,
+      "grad_norm": 0.3366181552410126,
+      "learning_rate": 0.0002,
+      "loss": 0.9297,
+      "step": 8230
+    },
+    {
+      "epoch": 0.5407711238720263,
+      "grad_norm": 0.34822607040405273,
+      "learning_rate": 0.0002,
+      "loss": 0.9439,
+      "step": 8240
+    },
+    {
+      "epoch": 0.5414273995077933,
+      "grad_norm": 0.3682822287082672,
+      "learning_rate": 0.0002,
+      "loss": 0.925,
+      "step": 8250
+    },
+    {
+      "epoch": 0.5420836751435603,
+      "grad_norm": 0.39051371812820435,
+      "learning_rate": 0.0002,
+      "loss": 0.9246,
+      "step": 8260
+    },
+    {
+      "epoch": 0.5427399507793274,
+      "grad_norm": 0.37731602787971497,
+      "learning_rate": 0.0002,
+      "loss": 0.9008,
+      "step": 8270
+    },
+    {
+      "epoch": 0.5433962264150943,
+      "grad_norm": 0.34222212433815,
+      "learning_rate": 0.0002,
+      "loss": 0.8849,
+      "step": 8280
+    },
+    {
+      "epoch": 0.5440525020508613,
+      "grad_norm": 0.3870106637477875,
+      "learning_rate": 0.0002,
+      "loss": 0.8956,
+      "step": 8290
+    },
+    {
+      "epoch": 0.5447087776866284,
+      "grad_norm": 0.35894039273262024,
+      "learning_rate": 0.0002,
+      "loss": 0.8974,
+      "step": 8300
+    },
+    {
+      "epoch": 0.5453650533223954,
+      "grad_norm": 0.38450875878334045,
+      "learning_rate": 0.0002,
+      "loss": 0.9664,
+      "step": 8310
+    },
+    {
+      "epoch": 0.5460213289581625,
+      "grad_norm": 0.3106869161128998,
+      "learning_rate": 0.0002,
+      "loss": 0.9024,
+      "step": 8320
+    },
+    {
+      "epoch": 0.5466776045939294,
+      "grad_norm": 0.35935118794441223,
+      "learning_rate": 0.0002,
+      "loss": 0.8578,
+      "step": 8330
+    },
+    {
+      "epoch": 0.5473338802296964,
+      "grad_norm": 0.37256261706352234,
+      "learning_rate": 0.0002,
+      "loss": 0.904,
+      "step": 8340
+    },
+    {
+      "epoch": 0.5479901558654635,
+      "grad_norm": 0.3807767629623413,
+      "learning_rate": 0.0002,
+      "loss": 0.9091,
+      "step": 8350
+    },
+    {
+      "epoch": 0.5486464315012305,
+      "grad_norm": 0.3551439046859741,
+      "learning_rate": 0.0002,
+      "loss": 0.9314,
+      "step": 8360
+    },
+    {
+      "epoch": 0.5493027071369976,
+      "grad_norm": 0.39783644676208496,
+      "learning_rate": 0.0002,
+      "loss": 0.9046,
+      "step": 8370
+    },
+    {
+      "epoch": 0.5499589827727646,
+      "grad_norm": 0.37729522585868835,
+      "learning_rate": 0.0002,
+      "loss": 0.9466,
+      "step": 8380
+    },
+    {
+      "epoch": 0.5506152584085315,
+      "grad_norm": 0.32453545928001404,
+      "learning_rate": 0.0002,
+      "loss": 0.9559,
+      "step": 8390
+    },
+    {
+      "epoch": 0.5512715340442986,
+      "grad_norm": 0.353837251663208,
+      "learning_rate": 0.0002,
+      "loss": 0.9073,
+      "step": 8400
+    },
+    {
+      "epoch": 0.5519278096800656,
+      "grad_norm": 0.37711212038993835,
+      "learning_rate": 0.0002,
+      "loss": 0.961,
+      "step": 8410
+    },
+    {
+      "epoch": 0.5525840853158327,
+      "grad_norm": 0.3966476619243622,
+      "learning_rate": 0.0002,
+      "loss": 0.9844,
+      "step": 8420
+    },
+    {
+      "epoch": 0.5532403609515997,
+      "grad_norm": 0.338074266910553,
+      "learning_rate": 0.0002,
+      "loss": 0.9221,
+      "step": 8430
+    },
+    {
+      "epoch": 0.5538966365873667,
+      "grad_norm": 0.3984934389591217,
+      "learning_rate": 0.0002,
+      "loss": 0.9101,
+      "step": 8440
+    },
+    {
+      "epoch": 0.5545529122231337,
+      "grad_norm": 0.3430469334125519,
+      "learning_rate": 0.0002,
+      "loss": 0.9177,
+      "step": 8450
+    },
+    {
+      "epoch": 0.5552091878589007,
+      "grad_norm": 0.34223586320877075,
+      "learning_rate": 0.0002,
+      "loss": 0.9128,
+      "step": 8460
+    },
+    {
+      "epoch": 0.5558654634946678,
+      "grad_norm": 0.4057091474533081,
+      "learning_rate": 0.0002,
+      "loss": 0.8961,
+      "step": 8470
+    },
+    {
+      "epoch": 0.5565217391304348,
+      "grad_norm": 0.39272257685661316,
+      "learning_rate": 0.0002,
+      "loss": 0.9121,
+      "step": 8480
+    },
+    {
+      "epoch": 0.5571780147662018,
+      "grad_norm": 0.34801942110061646,
+      "learning_rate": 0.0002,
+      "loss": 0.8941,
+      "step": 8490
+    },
+    {
+      "epoch": 0.5578342904019689,
+      "grad_norm": 0.5177333354949951,
+      "learning_rate": 0.0002,
+      "loss": 0.8771,
+      "step": 8500
+    },
+    {
+      "epoch": 0.5584905660377358,
+      "grad_norm": 0.4139311909675598,
+      "learning_rate": 0.0002,
+      "loss": 0.9289,
+      "step": 8510
+    },
+    {
+      "epoch": 0.5591468416735029,
+      "grad_norm": 0.3526647388935089,
+      "learning_rate": 0.0002,
+      "loss": 0.9391,
+      "step": 8520
+    },
+    {
+      "epoch": 0.5598031173092699,
+      "grad_norm": 0.3475663661956787,
+      "learning_rate": 0.0002,
+      "loss": 0.9354,
+      "step": 8530
+    },
+    {
+      "epoch": 0.5604593929450369,
+      "grad_norm": 0.3338899314403534,
+      "learning_rate": 0.0002,
+      "loss": 0.9623,
+      "step": 8540
+    },
+    {
+      "epoch": 0.561115668580804,
+      "grad_norm": 0.392212837934494,
+      "learning_rate": 0.0002,
+      "loss": 0.9156,
+      "step": 8550
+    },
+    {
+      "epoch": 0.561771944216571,
+      "grad_norm": 0.3058992028236389,
+      "learning_rate": 0.0002,
+      "loss": 0.9469,
+      "step": 8560
+    },
+    {
+      "epoch": 0.5624282198523379,
+      "grad_norm": 0.40603798627853394,
+      "learning_rate": 0.0002,
+      "loss": 0.8465,
+      "step": 8570
+    },
+    {
+      "epoch": 0.563084495488105,
+      "grad_norm": 0.37745383381843567,
+      "learning_rate": 0.0002,
+      "loss": 0.8962,
+      "step": 8580
+    },
+    {
+      "epoch": 0.563740771123872,
+      "grad_norm": 0.40886175632476807,
+      "learning_rate": 0.0002,
+      "loss": 0.8787,
+      "step": 8590
+    },
+    {
+      "epoch": 0.5643970467596391,
+      "grad_norm": 0.3653600215911865,
+      "learning_rate": 0.0002,
+      "loss": 0.9469,
+      "step": 8600
+    },
+    {
+      "epoch": 0.5650533223954061,
+      "grad_norm": 0.32329508662223816,
+      "learning_rate": 0.0002,
+      "loss": 0.8976,
+      "step": 8610
+    },
+    {
+      "epoch": 0.565709598031173,
+      "grad_norm": 0.3535236716270447,
+      "learning_rate": 0.0002,
+      "loss": 0.9081,
+      "step": 8620
+    },
+    {
+      "epoch": 0.5663658736669401,
+      "grad_norm": 0.37497273087501526,
+      "learning_rate": 0.0002,
+      "loss": 0.8703,
+      "step": 8630
+    },
+    {
+      "epoch": 0.5670221493027071,
+      "grad_norm": 0.3781719505786896,
+      "learning_rate": 0.0002,
+      "loss": 0.8853,
+      "step": 8640
+    },
+    {
+      "epoch": 0.5676784249384742,
+      "grad_norm": 0.4481894075870514,
+      "learning_rate": 0.0002,
+      "loss": 0.9074,
+      "step": 8650
+    },
+    {
+      "epoch": 0.5683347005742412,
+      "grad_norm": 0.35790109634399414,
+      "learning_rate": 0.0002,
+      "loss": 0.8923,
+      "step": 8660
+    },
+    {
+      "epoch": 0.5689909762100082,
+      "grad_norm": 0.4617280662059784,
+      "learning_rate": 0.0002,
+      "loss": 0.8946,
+      "step": 8670
+    },
+    {
+      "epoch": 0.5696472518457752,
+      "grad_norm": 0.3813382089138031,
+      "learning_rate": 0.0002,
+      "loss": 0.9059,
+      "step": 8680
+    },
+    {
+      "epoch": 0.5703035274815422,
+      "grad_norm": 0.33388257026672363,
+      "learning_rate": 0.0002,
+      "loss": 0.9126,
+      "step": 8690
+    },
+    {
+      "epoch": 0.5709598031173093,
+      "grad_norm": 0.330185204744339,
+      "learning_rate": 0.0002,
+      "loss": 0.8866,
+      "step": 8700
+    },
+    {
+      "epoch": 0.5716160787530763,
+      "grad_norm": 0.3768845200538635,
+      "learning_rate": 0.0002,
+      "loss": 0.916,
+      "step": 8710
+    },
+    {
+      "epoch": 0.5722723543888433,
+      "grad_norm": 0.3221369683742523,
+      "learning_rate": 0.0002,
+      "loss": 0.9213,
+      "step": 8720
+    },
+    {
+      "epoch": 0.5729286300246104,
+      "grad_norm": 0.3221580684185028,
+      "learning_rate": 0.0002,
+      "loss": 0.8919,
+      "step": 8730
+    },
+    {
+      "epoch": 0.5735849056603773,
+      "grad_norm": 0.3693002760410309,
+      "learning_rate": 0.0002,
+      "loss": 0.924,
+      "step": 8740
+    },
+    {
+      "epoch": 0.5742411812961444,
+      "grad_norm": 0.3859325349330902,
+      "learning_rate": 0.0002,
+      "loss": 0.8983,
+      "step": 8750
+    },
+    {
+      "epoch": 0.5748974569319114,
+      "grad_norm": 0.3891449570655823,
+      "learning_rate": 0.0002,
+      "loss": 0.9321,
+      "step": 8760
+    },
+    {
+      "epoch": 0.5755537325676784,
+      "grad_norm": 0.3313274681568146,
+      "learning_rate": 0.0002,
+      "loss": 0.9315,
+      "step": 8770
+    },
+    {
+      "epoch": 0.5762100082034455,
+      "grad_norm": 0.33418914675712585,
+      "learning_rate": 0.0002,
+      "loss": 0.89,
+      "step": 8780
+    },
+    {
+      "epoch": 0.5768662838392125,
+      "grad_norm": 0.4153108596801758,
+      "learning_rate": 0.0002,
+      "loss": 0.9029,
+      "step": 8790
+    },
+    {
+      "epoch": 0.5775225594749795,
+      "grad_norm": 0.3614528477191925,
+      "learning_rate": 0.0002,
+      "loss": 0.9126,
+      "step": 8800
+    },
+    {
+      "epoch": 0.5781788351107465,
+      "grad_norm": 0.3172459602355957,
+      "learning_rate": 0.0002,
+      "loss": 0.9086,
+      "step": 8810
+    },
+    {
+      "epoch": 0.5788351107465135,
+      "grad_norm": 0.3795868456363678,
+      "learning_rate": 0.0002,
+      "loss": 0.9225,
+      "step": 8820
+    },
+    {
+      "epoch": 0.5794913863822806,
+      "grad_norm": 0.3750050961971283,
+      "learning_rate": 0.0002,
+      "loss": 0.9199,
+      "step": 8830
+    },
+    {
+      "epoch": 0.5801476620180476,
+      "grad_norm": 0.3348597288131714,
+      "learning_rate": 0.0002,
+      "loss": 0.9527,
+      "step": 8840
+    },
+    {
+      "epoch": 0.5808039376538146,
+      "grad_norm": 0.39116451144218445,
+      "learning_rate": 0.0002,
+      "loss": 0.9379,
+      "step": 8850
+    },
+    {
+      "epoch": 0.5814602132895816,
+      "grad_norm": 0.33639633655548096,
+      "learning_rate": 0.0002,
+      "loss": 0.9087,
+      "step": 8860
+    },
+    {
+      "epoch": 0.5821164889253486,
+      "grad_norm": 0.33336129784584045,
+      "learning_rate": 0.0002,
+      "loss": 0.9181,
+      "step": 8870
+    },
+    {
+      "epoch": 0.5827727645611157,
+      "grad_norm": 0.3624817132949829,
+      "learning_rate": 0.0002,
+      "loss": 0.9194,
+      "step": 8880
+    },
+    {
+      "epoch": 0.5834290401968827,
+      "grad_norm": 0.2976662218570709,
+      "learning_rate": 0.0002,
+      "loss": 0.9059,
+      "step": 8890
+    },
+    {
+      "epoch": 0.5840853158326497,
+      "grad_norm": 0.4231838583946228,
+      "learning_rate": 0.0002,
+      "loss": 0.8903,
+      "step": 8900
+    },
+    {
+      "epoch": 0.5847415914684168,
+      "grad_norm": 0.343156099319458,
+      "learning_rate": 0.0002,
+      "loss": 0.9157,
+      "step": 8910
+    },
+    {
+      "epoch": 0.5853978671041837,
+      "grad_norm": 0.39717403054237366,
+      "learning_rate": 0.0002,
+      "loss": 0.9655,
+      "step": 8920
+    },
+    {
+      "epoch": 0.5860541427399508,
+      "grad_norm": 0.3388144075870514,
+      "learning_rate": 0.0002,
+      "loss": 0.8758,
+      "step": 8930
+    },
+    {
+      "epoch": 0.5867104183757178,
+      "grad_norm": 0.3949063718318939,
+      "learning_rate": 0.0002,
+      "loss": 0.922,
+      "step": 8940
+    },
+    {
+      "epoch": 0.5873666940114848,
+      "grad_norm": 0.358415424823761,
+      "learning_rate": 0.0002,
+      "loss": 0.9081,
+      "step": 8950
+    },
+    {
+      "epoch": 0.5880229696472519,
+      "grad_norm": 0.367883563041687,
+      "learning_rate": 0.0002,
+      "loss": 0.906,
+      "step": 8960
+    },
+    {
+      "epoch": 0.5886792452830188,
+      "grad_norm": 0.3859010636806488,
+      "learning_rate": 0.0002,
+      "loss": 0.9131,
+      "step": 8970
+    },
+    {
+      "epoch": 0.5893355209187859,
+      "grad_norm": 0.3540095090866089,
+      "learning_rate": 0.0002,
+      "loss": 0.916,
+      "step": 8980
+    },
+    {
+      "epoch": 0.5899917965545529,
+      "grad_norm": 0.33913302421569824,
+      "learning_rate": 0.0002,
+      "loss": 0.9704,
+      "step": 8990
+    },
+    {
+      "epoch": 0.5906480721903199,
+      "grad_norm": 0.35641804337501526,
+      "learning_rate": 0.0002,
+      "loss": 0.8772,
+      "step": 9000
+    },
+    {
+      "epoch": 0.591304347826087,
+      "grad_norm": 0.3448907434940338,
+      "learning_rate": 0.0002,
+      "loss": 0.8644,
+      "step": 9010
+    },
+    {
+      "epoch": 0.591960623461854,
+      "grad_norm": 0.4147050082683563,
+      "learning_rate": 0.0002,
+      "loss": 0.9059,
+      "step": 9020
+    },
+    {
+      "epoch": 0.592616899097621,
+      "grad_norm": 0.45032307505607605,
+      "learning_rate": 0.0002,
+      "loss": 0.9134,
+      "step": 9030
+    },
+    {
+      "epoch": 0.593273174733388,
+      "grad_norm": 0.3628511428833008,
+      "learning_rate": 0.0002,
+      "loss": 0.8855,
+      "step": 9040
+    },
+    {
+      "epoch": 0.593929450369155,
+      "grad_norm": 0.42760607600212097,
+      "learning_rate": 0.0002,
+      "loss": 0.8924,
+      "step": 9050
+    },
+    {
+      "epoch": 0.5945857260049221,
+      "grad_norm": 0.3591140806674957,
+      "learning_rate": 0.0002,
+      "loss": 0.889,
+      "step": 9060
+    },
+    {
+      "epoch": 0.5952420016406891,
+      "grad_norm": 0.3496510982513428,
+      "learning_rate": 0.0002,
+      "loss": 0.9409,
+      "step": 9070
+    },
+    {
+      "epoch": 0.5958982772764561,
+      "grad_norm": 0.4098506569862366,
+      "learning_rate": 0.0002,
+      "loss": 0.9394,
+      "step": 9080
+    },
+    {
+      "epoch": 0.5965545529122231,
+      "grad_norm": 0.3535410165786743,
+      "learning_rate": 0.0002,
+      "loss": 0.9455,
+      "step": 9090
+    },
+    {
+      "epoch": 0.5972108285479901,
+      "grad_norm": 0.3716369569301605,
+      "learning_rate": 0.0002,
+      "loss": 0.8814,
+      "step": 9100
+    },
+    {
+      "epoch": 0.5978671041837572,
+      "grad_norm": 0.3236614763736725,
+      "learning_rate": 0.0002,
+      "loss": 0.8656,
+      "step": 9110
+    },
+    {
+      "epoch": 0.5985233798195242,
+      "grad_norm": 0.311577171087265,
+      "learning_rate": 0.0002,
+      "loss": 0.905,
+      "step": 9120
+    },
+    {
+      "epoch": 0.5991796554552912,
+      "grad_norm": 0.3261461555957794,
+      "learning_rate": 0.0002,
+      "loss": 0.9118,
+      "step": 9130
+    },
+    {
+      "epoch": 0.5998359310910583,
+      "grad_norm": 0.3794492781162262,
+      "learning_rate": 0.0002,
+      "loss": 0.9274,
+      "step": 9140
+    },
+    {
+      "epoch": 0.6004922067268252,
+      "grad_norm": 0.31726357340812683,
+      "learning_rate": 0.0002,
+      "loss": 0.8967,
+      "step": 9150
+    },
+    {
+      "epoch": 0.6011484823625923,
+      "grad_norm": 0.38432174921035767,
+      "learning_rate": 0.0002,
+      "loss": 0.9156,
+      "step": 9160
+    },
+    {
+      "epoch": 0.6018047579983593,
+      "grad_norm": 0.4364495873451233,
+      "learning_rate": 0.0002,
+      "loss": 0.9703,
+      "step": 9170
+    },
+    {
+      "epoch": 0.6024610336341263,
+      "grad_norm": 0.41058987379074097,
+      "learning_rate": 0.0002,
+      "loss": 0.9162,
+      "step": 9180
+    },
+    {
+      "epoch": 0.6031173092698934,
+      "grad_norm": 0.3146302103996277,
+      "learning_rate": 0.0002,
+      "loss": 0.8914,
+      "step": 9190
+    },
+    {
+      "epoch": 0.6037735849056604,
+      "grad_norm": 0.3117610216140747,
+      "learning_rate": 0.0002,
+      "loss": 0.8859,
+      "step": 9200
+    },
+    {
+      "epoch": 0.6044298605414274,
+      "grad_norm": 0.31921523809432983,
+      "learning_rate": 0.0002,
+      "loss": 0.8741,
+      "step": 9210
+    },
+    {
+      "epoch": 0.6050861361771944,
+      "grad_norm": 0.3295772969722748,
+      "learning_rate": 0.0002,
+      "loss": 0.8756,
+      "step": 9220
+    },
+    {
+      "epoch": 0.6057424118129614,
+      "grad_norm": 0.3216910660266876,
+      "learning_rate": 0.0002,
+      "loss": 0.8578,
+      "step": 9230
+    },
+    {
+      "epoch": 0.6063986874487285,
+      "grad_norm": 0.40590721368789673,
+      "learning_rate": 0.0002,
+      "loss": 0.9341,
+      "step": 9240
+    },
+    {
+      "epoch": 0.6070549630844955,
+      "grad_norm": 0.4337029457092285,
+      "learning_rate": 0.0002,
+      "loss": 0.935,
+      "step": 9250
+    },
+    {
+      "epoch": 0.6077112387202626,
+      "grad_norm": 0.3434010148048401,
+      "learning_rate": 0.0002,
+      "loss": 0.8866,
+      "step": 9260
+    },
+    {
+      "epoch": 0.6083675143560295,
+      "grad_norm": 0.340589702129364,
+      "learning_rate": 0.0002,
+      "loss": 0.8692,
+      "step": 9270
+    },
+    {
+      "epoch": 0.6090237899917965,
+      "grad_norm": 0.33650949597358704,
+      "learning_rate": 0.0002,
+      "loss": 0.8741,
+      "step": 9280
+    },
+    {
+      "epoch": 0.6096800656275636,
+      "grad_norm": 0.4195605218410492,
+      "learning_rate": 0.0002,
+      "loss": 0.9396,
+      "step": 9290
+    },
+    {
+      "epoch": 0.6103363412633306,
+      "grad_norm": 0.444624125957489,
+      "learning_rate": 0.0002,
+      "loss": 0.8877,
+      "step": 9300
+    },
+    {
+      "epoch": 0.6109926168990977,
+      "grad_norm": 0.5425066947937012,
+      "learning_rate": 0.0002,
+      "loss": 0.8792,
+      "step": 9310
+    },
+    {
+      "epoch": 0.6116488925348647,
+      "grad_norm": 0.36225831508636475,
+      "learning_rate": 0.0002,
+      "loss": 0.865,
+      "step": 9320
+    },
+    {
+      "epoch": 0.6123051681706316,
+      "grad_norm": 0.3421785533428192,
+      "learning_rate": 0.0002,
+      "loss": 0.9439,
+      "step": 9330
+    },
+    {
+      "epoch": 0.6129614438063987,
+      "grad_norm": 0.3792729079723358,
+      "learning_rate": 0.0002,
+      "loss": 0.8842,
+      "step": 9340
+    },
+    {
+      "epoch": 0.6136177194421657,
+      "grad_norm": 0.41844120621681213,
+      "learning_rate": 0.0002,
+      "loss": 0.8791,
+      "step": 9350
+    },
+    {
+      "epoch": 0.6142739950779327,
+      "grad_norm": 0.3430991768836975,
+      "learning_rate": 0.0002,
+      "loss": 0.9091,
+      "step": 9360
+    },
+    {
+      "epoch": 0.6149302707136998,
+      "grad_norm": 0.3460402190685272,
+      "learning_rate": 0.0002,
+      "loss": 0.9165,
+      "step": 9370
+    },
+    {
+      "epoch": 0.6155865463494667,
+      "grad_norm": 0.3775254189968109,
+      "learning_rate": 0.0002,
+      "loss": 0.8571,
+      "step": 9380
+    },
+    {
+      "epoch": 0.6162428219852338,
+      "grad_norm": 0.3706645965576172,
+      "learning_rate": 0.0002,
+      "loss": 0.8915,
+      "step": 9390
+    },
+    {
+      "epoch": 0.6168990976210008,
+      "grad_norm": 0.3817055821418762,
+      "learning_rate": 0.0002,
+      "loss": 0.9085,
+      "step": 9400
+    },
+    {
+      "epoch": 0.6175553732567678,
+      "grad_norm": 0.3856641352176666,
+      "learning_rate": 0.0002,
+      "loss": 0.9086,
+      "step": 9410
+    },
+    {
+      "epoch": 0.6182116488925349,
+      "grad_norm": 0.3271346688270569,
+      "learning_rate": 0.0002,
+      "loss": 0.9031,
+      "step": 9420
+    },
+    {
+      "epoch": 0.6188679245283019,
+      "grad_norm": 0.334314227104187,
+      "learning_rate": 0.0002,
+      "loss": 0.883,
+      "step": 9430
+    },
+    {
+      "epoch": 0.619524200164069,
+      "grad_norm": 0.3850700557231903,
+      "learning_rate": 0.0002,
+      "loss": 0.9398,
+      "step": 9440
+    },
+    {
+      "epoch": 0.6201804757998359,
+      "grad_norm": 0.37492436170578003,
+      "learning_rate": 0.0002,
+      "loss": 0.9246,
+      "step": 9450
+    },
+    {
+      "epoch": 0.6208367514356029,
+      "grad_norm": 0.44262826442718506,
+      "learning_rate": 0.0002,
+      "loss": 0.9308,
+      "step": 9460
+    },
+    {
+      "epoch": 0.62149302707137,
+      "grad_norm": 0.4070657193660736,
+      "learning_rate": 0.0002,
+      "loss": 0.9659,
+      "step": 9470
+    },
+    {
+      "epoch": 0.622149302707137,
+      "grad_norm": 0.37636154890060425,
+      "learning_rate": 0.0002,
+      "loss": 0.8812,
+      "step": 9480
+    },
+    {
+      "epoch": 0.6228055783429041,
+      "grad_norm": 0.38988572359085083,
+      "learning_rate": 0.0002,
+      "loss": 0.941,
+      "step": 9490
+    },
+    {
+      "epoch": 0.623461853978671,
+      "grad_norm": 0.36479735374450684,
+      "learning_rate": 0.0002,
+      "loss": 0.8959,
+      "step": 9500
+    },
+    {
+      "epoch": 0.624118129614438,
+      "grad_norm": 0.3759172260761261,
+      "learning_rate": 0.0002,
+      "loss": 0.9431,
+      "step": 9510
+    },
+    {
+      "epoch": 0.6247744052502051,
+      "grad_norm": 0.3380950689315796,
+      "learning_rate": 0.0002,
+      "loss": 0.889,
+      "step": 9520
+    },
+    {
+      "epoch": 0.6254306808859721,
+      "grad_norm": 0.3483046889305115,
+      "learning_rate": 0.0002,
+      "loss": 0.8597,
+      "step": 9530
+    },
+    {
+      "epoch": 0.6260869565217392,
+      "grad_norm": 0.3562379777431488,
+      "learning_rate": 0.0002,
+      "loss": 0.9195,
+      "step": 9540
+    },
+    {
+      "epoch": 0.6267432321575062,
+      "grad_norm": 0.40167364478111267,
+      "learning_rate": 0.0002,
+      "loss": 0.9434,
+      "step": 9550
+    },
+    {
+      "epoch": 0.6273995077932731,
+      "grad_norm": 0.36348867416381836,
+      "learning_rate": 0.0002,
+      "loss": 0.8677,
+      "step": 9560
+    },
+    {
+      "epoch": 0.6280557834290402,
+      "grad_norm": 0.33701515197753906,
+      "learning_rate": 0.0002,
+      "loss": 0.9061,
+      "step": 9570
+    },
+    {
+      "epoch": 0.6287120590648072,
+      "grad_norm": 0.3795888125896454,
+      "learning_rate": 0.0002,
+      "loss": 0.8926,
+      "step": 9580
+    },
+    {
+      "epoch": 0.6293683347005743,
+      "grad_norm": 0.3701418340206146,
+      "learning_rate": 0.0002,
+      "loss": 0.8672,
+      "step": 9590
+    },
+    {
+      "epoch": 0.6300246103363413,
+      "grad_norm": 0.32559722661972046,
+      "learning_rate": 0.0002,
+      "loss": 0.9025,
+      "step": 9600
+    },
+    {
+      "epoch": 0.6306808859721083,
+      "grad_norm": 0.3093271553516388,
+      "learning_rate": 0.0002,
+      "loss": 0.9223,
+      "step": 9610
+    },
+    {
+      "epoch": 0.6313371616078753,
+      "grad_norm": 0.41225478053092957,
+      "learning_rate": 0.0002,
+      "loss": 0.9032,
+      "step": 9620
+    },
+    {
+      "epoch": 0.6319934372436423,
+      "grad_norm": 0.3798231780529022,
+      "learning_rate": 0.0002,
+      "loss": 0.9125,
+      "step": 9630
+    },
+    {
+      "epoch": 0.6326497128794093,
+      "grad_norm": 0.38690295815467834,
+      "learning_rate": 0.0002,
+      "loss": 0.9252,
+      "step": 9640
+    },
+    {
+      "epoch": 0.6333059885151764,
+      "grad_norm": 0.3629007637500763,
+      "learning_rate": 0.0002,
+      "loss": 0.8668,
+      "step": 9650
+    },
+    {
+      "epoch": 0.6339622641509434,
+      "grad_norm": 0.39467132091522217,
+      "learning_rate": 0.0002,
+      "loss": 0.9143,
+      "step": 9660
+    },
+    {
+      "epoch": 0.6346185397867105,
+      "grad_norm": 0.37746182084083557,
+      "learning_rate": 0.0002,
+      "loss": 0.9267,
+      "step": 9670
+    },
+    {
+      "epoch": 0.6352748154224774,
+      "grad_norm": 0.3718436062335968,
+      "learning_rate": 0.0002,
+      "loss": 0.8915,
+      "step": 9680
+    },
+    {
+      "epoch": 0.6359310910582444,
+      "grad_norm": 0.3951144218444824,
+      "learning_rate": 0.0002,
+      "loss": 0.9111,
+      "step": 9690
+    },
+    {
+      "epoch": 0.6365873666940115,
+      "grad_norm": 0.4104543924331665,
+      "learning_rate": 0.0002,
+      "loss": 0.897,
+      "step": 9700
+    },
+    {
+      "epoch": 0.6372436423297785,
+      "grad_norm": 0.37318137288093567,
+      "learning_rate": 0.0002,
+      "loss": 0.9349,
+      "step": 9710
+    },
+    {
+      "epoch": 0.6378999179655456,
+      "grad_norm": 0.35614442825317383,
+      "learning_rate": 0.0002,
+      "loss": 0.9009,
+      "step": 9720
+    },
+    {
+      "epoch": 0.6385561936013125,
+      "grad_norm": 0.37922942638397217,
+      "learning_rate": 0.0002,
+      "loss": 0.9858,
+      "step": 9730
+    },
+    {
+      "epoch": 0.6392124692370795,
+      "grad_norm": 0.35624784231185913,
+      "learning_rate": 0.0002,
+      "loss": 0.8533,
+      "step": 9740
+    },
+    {
+      "epoch": 0.6398687448728466,
+      "grad_norm": 0.34946876764297485,
+      "learning_rate": 0.0002,
+      "loss": 0.934,
+      "step": 9750
+    },
+    {
+      "epoch": 0.6405250205086136,
+      "grad_norm": 0.37666648626327515,
+      "learning_rate": 0.0002,
+      "loss": 0.9176,
+      "step": 9760
+    },
+    {
+      "epoch": 0.6411812961443807,
+      "grad_norm": 0.3619046211242676,
+      "learning_rate": 0.0002,
+      "loss": 0.9297,
+      "step": 9770
+    },
+    {
+      "epoch": 0.6418375717801477,
+      "grad_norm": 0.3814936578273773,
+      "learning_rate": 0.0002,
+      "loss": 0.9135,
+      "step": 9780
+    },
+    {
+      "epoch": 0.6424938474159146,
+      "grad_norm": 0.4181577265262604,
+      "learning_rate": 0.0002,
+      "loss": 0.9143,
+      "step": 9790
+    },
+    {
+      "epoch": 0.6431501230516817,
+      "grad_norm": 0.3707144558429718,
+      "learning_rate": 0.0002,
+      "loss": 0.8978,
+      "step": 9800
+    },
+    {
+      "epoch": 0.6438063986874487,
+      "grad_norm": 0.3364716172218323,
+      "learning_rate": 0.0002,
+      "loss": 0.9311,
+      "step": 9810
+    },
+    {
+      "epoch": 0.6444626743232158,
+      "grad_norm": 0.31425684690475464,
+      "learning_rate": 0.0002,
+      "loss": 0.8566,
+      "step": 9820
+    },
+    {
+      "epoch": 0.6451189499589828,
+      "grad_norm": 0.32928282022476196,
+      "learning_rate": 0.0002,
+      "loss": 0.8792,
+      "step": 9830
+    },
+    {
+      "epoch": 0.6457752255947498,
+      "grad_norm": 0.39178264141082764,
+      "learning_rate": 0.0002,
+      "loss": 0.9232,
+      "step": 9840
+    },
+    {
+      "epoch": 0.6464315012305168,
+      "grad_norm": 0.37753361463546753,
+      "learning_rate": 0.0002,
+      "loss": 0.9413,
+      "step": 9850
+    },
+    {
+      "epoch": 0.6470877768662838,
+      "grad_norm": 0.4574730396270752,
+      "learning_rate": 0.0002,
+      "loss": 0.8777,
+      "step": 9860
+    },
+    {
+      "epoch": 0.6477440525020509,
+      "grad_norm": 0.3926962614059448,
+      "learning_rate": 0.0002,
+      "loss": 0.9078,
+      "step": 9870
+    },
+    {
+      "epoch": 0.6484003281378179,
+      "grad_norm": 0.3919081687927246,
+      "learning_rate": 0.0002,
+      "loss": 0.9221,
+      "step": 9880
+    },
+    {
+      "epoch": 0.6490566037735849,
+      "grad_norm": 0.3904387652873993,
+      "learning_rate": 0.0002,
+      "loss": 0.8681,
+      "step": 9890
+    },
+    {
+      "epoch": 0.649712879409352,
+      "grad_norm": 0.40808236598968506,
+      "learning_rate": 0.0002,
+      "loss": 0.883,
+      "step": 9900
+    },
+    {
+      "epoch": 0.6503691550451189,
+      "grad_norm": 0.40531080961227417,
+      "learning_rate": 0.0002,
+      "loss": 0.9133,
+      "step": 9910
+    },
+    {
+      "epoch": 0.6510254306808859,
+      "grad_norm": 0.3374323844909668,
+      "learning_rate": 0.0002,
+      "loss": 0.9417,
+      "step": 9920
+    },
+    {
+      "epoch": 0.651681706316653,
+      "grad_norm": 0.4000678062438965,
+      "learning_rate": 0.0002,
+      "loss": 0.9769,
+      "step": 9930
+    },
+    {
+      "epoch": 0.65233798195242,
+      "grad_norm": 0.35415270924568176,
+      "learning_rate": 0.0002,
+      "loss": 0.9069,
+      "step": 9940
+    },
+    {
+      "epoch": 0.6529942575881871,
+      "grad_norm": 0.3612231910228729,
+      "learning_rate": 0.0002,
+      "loss": 0.9337,
+      "step": 9950
+    },
+    {
+      "epoch": 0.6536505332239541,
+      "grad_norm": 0.3081146776676178,
+      "learning_rate": 0.0002,
+      "loss": 0.9402,
+      "step": 9960
+    },
+    {
+      "epoch": 0.654306808859721,
+      "grad_norm": 0.31079018115997314,
+      "learning_rate": 0.0002,
+      "loss": 0.8942,
+      "step": 9970
+    },
+    {
+      "epoch": 0.6549630844954881,
+      "grad_norm": 0.4256346523761749,
+      "learning_rate": 0.0002,
+      "loss": 0.905,
+      "step": 9980
+    },
+    {
+      "epoch": 0.6556193601312551,
+      "grad_norm": 0.3592916429042816,
+      "learning_rate": 0.0002,
+      "loss": 0.9226,
+      "step": 9990
+    },
+    {
+      "epoch": 0.6562756357670222,
+      "grad_norm": 0.3496004641056061,
+      "learning_rate": 0.0002,
+      "loss": 0.9018,
+      "step": 10000
+    },
+    {
+      "epoch": 0.6569319114027892,
+      "grad_norm": 0.4177037179470062,
+      "learning_rate": 0.0002,
+      "loss": 0.931,
+      "step": 10010
+    },
+    {
+      "epoch": 0.6575881870385561,
+      "grad_norm": 0.36680638790130615,
+      "learning_rate": 0.0002,
+      "loss": 0.9034,
+      "step": 10020
+    },
+    {
+      "epoch": 0.6582444626743232,
+      "grad_norm": 0.3916943073272705,
+      "learning_rate": 0.0002,
+      "loss": 0.9706,
+      "step": 10030
+    },
+    {
+      "epoch": 0.6589007383100902,
+      "grad_norm": 0.44537290930747986,
+      "learning_rate": 0.0002,
+      "loss": 0.886,
+      "step": 10040
+    },
+    {
+      "epoch": 0.6595570139458573,
+      "grad_norm": 0.3404697775840759,
+      "learning_rate": 0.0002,
+      "loss": 0.8798,
+      "step": 10050
+    },
+    {
+      "epoch": 0.6602132895816243,
+      "grad_norm": 0.3304594159126282,
+      "learning_rate": 0.0002,
+      "loss": 0.9801,
+      "step": 10060
+    },
+    {
+      "epoch": 0.6608695652173913,
+      "grad_norm": 0.39067313075065613,
+      "learning_rate": 0.0002,
+      "loss": 0.9107,
+      "step": 10070
+    },
+    {
+      "epoch": 0.6615258408531584,
+      "grad_norm": 0.366178959608078,
+      "learning_rate": 0.0002,
+      "loss": 0.9336,
+      "step": 10080
+    },
+    {
+      "epoch": 0.6621821164889253,
+      "grad_norm": 0.3880734145641327,
+      "learning_rate": 0.0002,
+      "loss": 0.8817,
+      "step": 10090
+    },
+    {
+      "epoch": 0.6628383921246924,
+      "grad_norm": 0.42047396302223206,
+      "learning_rate": 0.0002,
+      "loss": 0.9087,
+      "step": 10100
+    },
+    {
+      "epoch": 0.6634946677604594,
+      "grad_norm": 0.3299349546432495,
+      "learning_rate": 0.0002,
+      "loss": 0.8859,
+      "step": 10110
+    },
+    {
+      "epoch": 0.6641509433962264,
+      "grad_norm": 0.33470937609672546,
+      "learning_rate": 0.0002,
+      "loss": 0.9461,
+      "step": 10120
+    },
+    {
+      "epoch": 0.6648072190319935,
+      "grad_norm": 0.36240577697753906,
+      "learning_rate": 0.0002,
+      "loss": 0.9453,
+      "step": 10130
+    },
+    {
+      "epoch": 0.6654634946677604,
+      "grad_norm": 0.41457104682922363,
+      "learning_rate": 0.0002,
+      "loss": 0.8706,
+      "step": 10140
+    },
+    {
+      "epoch": 0.6661197703035275,
+      "grad_norm": 0.39064788818359375,
+      "learning_rate": 0.0002,
+      "loss": 0.9385,
+      "step": 10150
+    },
+    {
+      "epoch": 0.6667760459392945,
+      "grad_norm": 0.4137183725833893,
+      "learning_rate": 0.0002,
+      "loss": 0.9252,
+      "step": 10160
+    },
+    {
+      "epoch": 0.6674323215750615,
+      "grad_norm": 0.363413542509079,
+      "learning_rate": 0.0002,
+      "loss": 0.8874,
+      "step": 10170
+    },
+    {
+      "epoch": 0.6680885972108286,
+      "grad_norm": 0.3631424307823181,
+      "learning_rate": 0.0002,
+      "loss": 0.8936,
+      "step": 10180
+    },
+    {
+      "epoch": 0.6687448728465956,
+      "grad_norm": 0.460721880197525,
+      "learning_rate": 0.0002,
+      "loss": 0.9181,
+      "step": 10190
+    },
+    {
+      "epoch": 0.6694011484823625,
+      "grad_norm": 0.3525084853172302,
+      "learning_rate": 0.0002,
+      "loss": 0.8879,
+      "step": 10200
+    },
+    {
+      "epoch": 0.6700574241181296,
+      "grad_norm": 0.34321045875549316,
+      "learning_rate": 0.0002,
+      "loss": 0.934,
+      "step": 10210
+    },
+    {
+      "epoch": 0.6707136997538966,
+      "grad_norm": 0.3397759795188904,
+      "learning_rate": 0.0002,
+      "loss": 0.9056,
+      "step": 10220
+    },
+    {
+      "epoch": 0.6713699753896637,
+      "grad_norm": 0.3680257201194763,
+      "learning_rate": 0.0002,
+      "loss": 0.8883,
+      "step": 10230
+    },
+    {
+      "epoch": 0.6720262510254307,
+      "grad_norm": 0.4023214876651764,
+      "learning_rate": 0.0002,
+      "loss": 0.9316,
+      "step": 10240
+    },
+    {
+      "epoch": 0.6726825266611977,
+      "grad_norm": 0.3645709753036499,
+      "learning_rate": 0.0002,
+      "loss": 0.8809,
+      "step": 10250
+    },
+    {
+      "epoch": 0.6733388022969647,
+      "grad_norm": 0.3558615744113922,
+      "learning_rate": 0.0002,
+      "loss": 0.9766,
+      "step": 10260
+    },
+    {
+      "epoch": 0.6739950779327317,
+      "grad_norm": 0.3725513517856598,
+      "learning_rate": 0.0002,
+      "loss": 0.872,
+      "step": 10270
+    },
+    {
+      "epoch": 0.6746513535684988,
+      "grad_norm": 0.42790961265563965,
+      "learning_rate": 0.0002,
+      "loss": 0.9354,
+      "step": 10280
+    },
+    {
+      "epoch": 0.6753076292042658,
+      "grad_norm": 0.3808377683162689,
+      "learning_rate": 0.0002,
+      "loss": 0.8955,
+      "step": 10290
+    },
+    {
+      "epoch": 0.6759639048400328,
+      "grad_norm": 0.40500468015670776,
+      "learning_rate": 0.0002,
+      "loss": 0.8761,
+      "step": 10300
+    },
+    {
+      "epoch": 0.6766201804757999,
+      "grad_norm": 0.3631184995174408,
+      "learning_rate": 0.0002,
+      "loss": 0.9212,
+      "step": 10310
+    },
+    {
+      "epoch": 0.6772764561115668,
+      "grad_norm": 0.3319573998451233,
+      "learning_rate": 0.0002,
+      "loss": 0.8867,
+      "step": 10320
+    },
+    {
+      "epoch": 0.6779327317473339,
+      "grad_norm": 0.3851188123226166,
+      "learning_rate": 0.0002,
+      "loss": 0.857,
+      "step": 10330
+    },
+    {
+      "epoch": 0.6785890073831009,
+      "grad_norm": 0.37413159012794495,
+      "learning_rate": 0.0002,
+      "loss": 0.9159,
+      "step": 10340
+    },
+    {
+      "epoch": 0.6792452830188679,
+      "grad_norm": 0.34124433994293213,
+      "learning_rate": 0.0002,
+      "loss": 0.8728,
+      "step": 10350
+    },
+    {
+      "epoch": 0.679901558654635,
+      "grad_norm": 0.3786381781101227,
+      "learning_rate": 0.0002,
+      "loss": 0.9654,
+      "step": 10360
+    },
+    {
+      "epoch": 0.680557834290402,
+      "grad_norm": 0.3848523199558258,
+      "learning_rate": 0.0002,
+      "loss": 0.915,
+      "step": 10370
+    },
+    {
+      "epoch": 0.681214109926169,
+      "grad_norm": 0.29365262389183044,
+      "learning_rate": 0.0002,
+      "loss": 0.8873,
+      "step": 10380
+    },
+    {
+      "epoch": 0.681870385561936,
+      "grad_norm": 0.3481557369232178,
+      "learning_rate": 0.0002,
+      "loss": 0.8692,
+      "step": 10390
+    },
+    {
+      "epoch": 0.682526661197703,
+      "grad_norm": 0.38596320152282715,
+      "learning_rate": 0.0002,
+      "loss": 0.885,
+      "step": 10400
+    },
+    {
+      "epoch": 0.6831829368334701,
+      "grad_norm": 0.30872032046318054,
+      "learning_rate": 0.0002,
+      "loss": 0.8545,
+      "step": 10410
+    },
+    {
+      "epoch": 0.6838392124692371,
+      "grad_norm": 0.3729351758956909,
+      "learning_rate": 0.0002,
+      "loss": 0.867,
+      "step": 10420
+    },
+    {
+      "epoch": 0.684495488105004,
+      "grad_norm": 0.35829052329063416,
+      "learning_rate": 0.0002,
+      "loss": 0.9089,
+      "step": 10430
+    },
+    {
+      "epoch": 0.6851517637407711,
+      "grad_norm": 0.34916967153549194,
+      "learning_rate": 0.0002,
+      "loss": 0.8797,
+      "step": 10440
+    },
+    {
+      "epoch": 0.6858080393765381,
+      "grad_norm": 0.40377020835876465,
+      "learning_rate": 0.0002,
+      "loss": 0.8731,
+      "step": 10450
+    },
+    {
+      "epoch": 0.6864643150123052,
+      "grad_norm": 0.36841881275177,
+      "learning_rate": 0.0002,
+      "loss": 0.9027,
+      "step": 10460
+    },
+    {
+      "epoch": 0.6871205906480722,
+      "grad_norm": 0.35462167859077454,
+      "learning_rate": 0.0002,
+      "loss": 0.8808,
+      "step": 10470
+    },
+    {
+      "epoch": 0.6877768662838392,
+      "grad_norm": 0.4630918502807617,
+      "learning_rate": 0.0002,
+      "loss": 0.967,
+      "step": 10480
+    },
+    {
+      "epoch": 0.6884331419196063,
+      "grad_norm": 0.4230295419692993,
+      "learning_rate": 0.0002,
+      "loss": 0.9114,
+      "step": 10490
+    },
+    {
+      "epoch": 0.6890894175553732,
+      "grad_norm": 0.35059425234794617,
+      "learning_rate": 0.0002,
+      "loss": 0.8908,
+      "step": 10500
+    },
+    {
+      "epoch": 0.6897456931911403,
+      "grad_norm": 0.4072548747062683,
+      "learning_rate": 0.0002,
+      "loss": 0.9325,
+      "step": 10510
+    },
+    {
+      "epoch": 0.6904019688269073,
+      "grad_norm": 0.33755314350128174,
+      "learning_rate": 0.0002,
+      "loss": 0.9026,
+      "step": 10520
+    },
+    {
+      "epoch": 0.6910582444626743,
+      "grad_norm": 0.4411669075489044,
+      "learning_rate": 0.0002,
+      "loss": 0.908,
+      "step": 10530
+    },
+    {
+      "epoch": 0.6917145200984414,
+      "grad_norm": 0.9155740737915039,
+      "learning_rate": 0.0002,
+      "loss": 0.8639,
+      "step": 10540
+    },
+    {
+      "epoch": 0.6923707957342083,
+      "grad_norm": 0.39329391717910767,
+      "learning_rate": 0.0002,
+      "loss": 0.8975,
+      "step": 10550
+    },
+    {
+      "epoch": 0.6930270713699754,
+      "grad_norm": 0.3602689206600189,
+      "learning_rate": 0.0002,
+      "loss": 0.92,
+      "step": 10560
+    },
+    {
+      "epoch": 0.6936833470057424,
+      "grad_norm": 0.35848474502563477,
+      "learning_rate": 0.0002,
+      "loss": 0.8247,
+      "step": 10570
+    },
+    {
+      "epoch": 0.6943396226415094,
+      "grad_norm": 0.37857621908187866,
+      "learning_rate": 0.0002,
+      "loss": 0.9171,
+      "step": 10580
+    },
+    {
+      "epoch": 0.6949958982772765,
+      "grad_norm": 0.37071332335472107,
+      "learning_rate": 0.0002,
+      "loss": 0.9533,
+      "step": 10590
+    },
+    {
+      "epoch": 0.6956521739130435,
+      "grad_norm": 0.33589425683021545,
+      "learning_rate": 0.0002,
+      "loss": 0.9237,
+      "step": 10600
+    },
+    {
+      "epoch": 0.6963084495488105,
+      "grad_norm": 0.4279285669326782,
+      "learning_rate": 0.0002,
+      "loss": 0.8772,
+      "step": 10610
+    },
+    {
+      "epoch": 0.6969647251845775,
+      "grad_norm": 0.47047463059425354,
+      "learning_rate": 0.0002,
+      "loss": 0.9452,
+      "step": 10620
+    },
+    {
+      "epoch": 0.6976210008203445,
+      "grad_norm": 0.41214805841445923,
+      "learning_rate": 0.0002,
+      "loss": 0.8996,
+      "step": 10630
+    },
+    {
+      "epoch": 0.6982772764561116,
+      "grad_norm": 0.3999953866004944,
+      "learning_rate": 0.0002,
+      "loss": 0.8776,
+      "step": 10640
+    },
+    {
+      "epoch": 0.6989335520918786,
+      "grad_norm": 0.42441290616989136,
+      "learning_rate": 0.0002,
+      "loss": 0.9426,
+      "step": 10650
+    },
+    {
+      "epoch": 0.6995898277276457,
+      "grad_norm": 0.41990748047828674,
+      "learning_rate": 0.0002,
+      "loss": 0.9319,
+      "step": 10660
+    },
+    {
+      "epoch": 0.7002461033634126,
+      "grad_norm": 0.37445810437202454,
+      "learning_rate": 0.0002,
+      "loss": 0.9024,
+      "step": 10670
+    },
+    {
+      "epoch": 0.7009023789991796,
+      "grad_norm": 0.37638649344444275,
+      "learning_rate": 0.0002,
+      "loss": 0.9029,
+      "step": 10680
+    },
+    {
+      "epoch": 0.7015586546349467,
+      "grad_norm": 0.41475534439086914,
+      "learning_rate": 0.0002,
+      "loss": 0.9211,
+      "step": 10690
+    },
+    {
+      "epoch": 0.7022149302707137,
+      "grad_norm": 0.3430996537208557,
+      "learning_rate": 0.0002,
+      "loss": 0.914,
+      "step": 10700
+    },
+    {
+      "epoch": 0.7028712059064807,
+      "grad_norm": 0.3569522798061371,
+      "learning_rate": 0.0002,
+      "loss": 0.9051,
+      "step": 10710
+    },
+    {
+      "epoch": 0.7035274815422478,
+      "grad_norm": 0.3145142197608948,
+      "learning_rate": 0.0002,
+      "loss": 0.8792,
+      "step": 10720
+    },
+    {
+      "epoch": 0.7041837571780147,
+      "grad_norm": 0.40700432658195496,
+      "learning_rate": 0.0002,
+      "loss": 0.9537,
+      "step": 10730
+    },
+    {
+      "epoch": 0.7048400328137818,
+      "grad_norm": 0.37982651591300964,
+      "learning_rate": 0.0002,
+      "loss": 0.8677,
+      "step": 10740
+    },
+    {
+      "epoch": 0.7054963084495488,
+      "grad_norm": 0.39240679144859314,
+      "learning_rate": 0.0002,
+      "loss": 0.8899,
+      "step": 10750
+    },
+    {
+      "epoch": 0.7061525840853158,
+      "grad_norm": 0.41788724064826965,
+      "learning_rate": 0.0002,
+      "loss": 0.9516,
+      "step": 10760
+    },
+    {
+      "epoch": 0.7068088597210829,
+      "grad_norm": 0.4015905559062958,
+      "learning_rate": 0.0002,
+      "loss": 0.8672,
+      "step": 10770
+    },
+    {
+      "epoch": 0.7074651353568499,
+      "grad_norm": 0.4019724130630493,
+      "learning_rate": 0.0002,
+      "loss": 0.9131,
+      "step": 10780
+    },
+    {
+      "epoch": 0.7081214109926169,
+      "grad_norm": 0.37387898564338684,
+      "learning_rate": 0.0002,
+      "loss": 0.8698,
+      "step": 10790
+    },
+    {
+      "epoch": 0.7087776866283839,
+      "grad_norm": 0.32087528705596924,
+      "learning_rate": 0.0002,
+      "loss": 0.8547,
+      "step": 10800
+    },
+    {
+      "epoch": 0.7094339622641509,
+      "grad_norm": 0.38160258531570435,
+      "learning_rate": 0.0002,
+      "loss": 0.9241,
+      "step": 10810
+    },
+    {
+      "epoch": 0.710090237899918,
+      "grad_norm": 0.340973436832428,
+      "learning_rate": 0.0002,
+      "loss": 0.8941,
+      "step": 10820
+    },
+    {
+      "epoch": 0.710746513535685,
+      "grad_norm": 0.3848867416381836,
+      "learning_rate": 0.0002,
+      "loss": 0.9404,
+      "step": 10830
+    },
+    {
+      "epoch": 0.711402789171452,
+      "grad_norm": 0.36736220121383667,
+      "learning_rate": 0.0002,
+      "loss": 0.9299,
+      "step": 10840
+    },
+    {
+      "epoch": 0.712059064807219,
+      "grad_norm": 0.367404967546463,
+      "learning_rate": 0.0002,
+      "loss": 0.8881,
+      "step": 10850
+    },
+    {
+      "epoch": 0.712715340442986,
+      "grad_norm": 0.333751380443573,
+      "learning_rate": 0.0002,
+      "loss": 0.9224,
+      "step": 10860
+    },
+    {
+      "epoch": 0.7133716160787531,
+      "grad_norm": 0.3865894079208374,
+      "learning_rate": 0.0002,
+      "loss": 0.8712,
+      "step": 10870
+    },
+    {
+      "epoch": 0.7140278917145201,
+      "grad_norm": 0.34016433358192444,
+      "learning_rate": 0.0002,
+      "loss": 0.8957,
+      "step": 10880
+    },
+    {
+      "epoch": 0.7146841673502872,
+      "grad_norm": 0.3233864903450012,
+      "learning_rate": 0.0002,
+      "loss": 0.8493,
+      "step": 10890
+    },
+    {
+      "epoch": 0.7153404429860541,
+      "grad_norm": 0.4275553822517395,
+      "learning_rate": 0.0002,
+      "loss": 0.9148,
+      "step": 10900
+    },
+    {
+      "epoch": 0.7159967186218211,
+      "grad_norm": 0.41715168952941895,
+      "learning_rate": 0.0002,
+      "loss": 0.9345,
+      "step": 10910
+    },
+    {
+      "epoch": 0.7166529942575882,
+      "grad_norm": 0.40540871024131775,
+      "learning_rate": 0.0002,
+      "loss": 0.9321,
+      "step": 10920
+    },
+    {
+      "epoch": 0.7173092698933552,
+      "grad_norm": 0.35520824790000916,
+      "learning_rate": 0.0002,
+      "loss": 0.9035,
+      "step": 10930
+    },
+    {
+      "epoch": 0.7179655455291223,
+      "grad_norm": 0.3805985748767853,
+      "learning_rate": 0.0002,
+      "loss": 0.9243,
+      "step": 10940
+    },
+    {
+      "epoch": 0.7186218211648893,
+      "grad_norm": 0.36969226598739624,
+      "learning_rate": 0.0002,
+      "loss": 0.9542,
+      "step": 10950
+    },
+    {
+      "epoch": 0.7192780968006562,
+      "grad_norm": 0.38408684730529785,
+      "learning_rate": 0.0002,
+      "loss": 0.892,
+      "step": 10960
+    },
+    {
+      "epoch": 0.7199343724364233,
+      "grad_norm": 0.3865699768066406,
+      "learning_rate": 0.0002,
+      "loss": 0.8699,
+      "step": 10970
+    },
+    {
+      "epoch": 0.7205906480721903,
+      "grad_norm": 0.37172383069992065,
+      "learning_rate": 0.0002,
+      "loss": 0.8767,
+      "step": 10980
+    },
+    {
+      "epoch": 0.7212469237079573,
+      "grad_norm": 0.39816758036613464,
+      "learning_rate": 0.0002,
+      "loss": 0.8609,
+      "step": 10990
+    },
+    {
+      "epoch": 0.7219031993437244,
+      "grad_norm": 0.4290051758289337,
+      "learning_rate": 0.0002,
+      "loss": 0.94,
+      "step": 11000
+    },
+    {
+      "epoch": 0.7225594749794914,
+      "grad_norm": 0.3733605146408081,
+      "learning_rate": 0.0002,
+      "loss": 0.9328,
+      "step": 11010
+    },
+    {
+      "epoch": 0.7232157506152584,
+      "grad_norm": 0.37305396795272827,
+      "learning_rate": 0.0002,
+      "loss": 0.9048,
+      "step": 11020
+    },
+    {
+      "epoch": 0.7238720262510254,
+      "grad_norm": 0.3557985723018646,
+      "learning_rate": 0.0002,
+      "loss": 0.8713,
+      "step": 11030
+    },
+    {
+      "epoch": 0.7245283018867924,
+      "grad_norm": 0.4635949730873108,
+      "learning_rate": 0.0002,
+      "loss": 0.8558,
+      "step": 11040
+    },
+    {
+      "epoch": 0.7251845775225595,
+      "grad_norm": 0.3419910967350006,
+      "learning_rate": 0.0002,
+      "loss": 0.8711,
+      "step": 11050
+    },
+    {
+      "epoch": 0.7258408531583265,
+      "grad_norm": 0.41529953479766846,
+      "learning_rate": 0.0002,
+      "loss": 0.9138,
+      "step": 11060
+    },
+    {
+      "epoch": 0.7264971287940936,
+      "grad_norm": 0.325923353433609,
+      "learning_rate": 0.0002,
+      "loss": 0.8859,
+      "step": 11070
+    },
+    {
+      "epoch": 0.7271534044298605,
+      "grad_norm": 0.41932010650634766,
+      "learning_rate": 0.0002,
+      "loss": 0.9924,
+      "step": 11080
+    },
+    {
+      "epoch": 0.7278096800656275,
+      "grad_norm": 0.39984336495399475,
+      "learning_rate": 0.0002,
+      "loss": 0.9662,
+      "step": 11090
+    },
+    {
+      "epoch": 0.7284659557013946,
+      "grad_norm": 0.3673977851867676,
+      "learning_rate": 0.0002,
+      "loss": 0.8849,
+      "step": 11100
+    },
+    {
+      "epoch": 0.7291222313371616,
+      "grad_norm": 0.3722262680530548,
+      "learning_rate": 0.0002,
+      "loss": 0.9201,
+      "step": 11110
+    },
+    {
+      "epoch": 0.7297785069729287,
+      "grad_norm": 0.3859870135784149,
+      "learning_rate": 0.0002,
+      "loss": 0.9156,
+      "step": 11120
+    },
+    {
+      "epoch": 0.7304347826086957,
+      "grad_norm": 0.3323642313480377,
+      "learning_rate": 0.0002,
+      "loss": 0.899,
+      "step": 11130
+    },
+    {
+      "epoch": 0.7310910582444626,
+      "grad_norm": 0.37793564796447754,
+      "learning_rate": 0.0002,
+      "loss": 0.9244,
+      "step": 11140
+    },
+    {
+      "epoch": 0.7317473338802297,
+      "grad_norm": 0.44874733686447144,
+      "learning_rate": 0.0002,
+      "loss": 0.8767,
+      "step": 11150
+    },
+    {
+      "epoch": 0.7324036095159967,
+      "grad_norm": 0.4104187786579132,
+      "learning_rate": 0.0002,
+      "loss": 0.9078,
+      "step": 11160
+    },
+    {
+      "epoch": 0.7330598851517638,
+      "grad_norm": 0.3699926435947418,
+      "learning_rate": 0.0002,
+      "loss": 0.8828,
+      "step": 11170
+    },
+    {
+      "epoch": 0.7337161607875308,
+      "grad_norm": 0.41013723611831665,
+      "learning_rate": 0.0002,
+      "loss": 0.8909,
+      "step": 11180
+    },
+    {
+      "epoch": 0.7343724364232977,
+      "grad_norm": 0.370669424533844,
+      "learning_rate": 0.0002,
+      "loss": 0.8934,
+      "step": 11190
+    },
+    {
+      "epoch": 0.7350287120590648,
+      "grad_norm": 0.4059436619281769,
+      "learning_rate": 0.0002,
+      "loss": 0.881,
+      "step": 11200
+    },
+    {
+      "epoch": 0.7356849876948318,
+      "grad_norm": 0.5038959383964539,
+      "learning_rate": 0.0002,
+      "loss": 0.903,
+      "step": 11210
+    },
+    {
+      "epoch": 0.7363412633305989,
+      "grad_norm": 0.337137371301651,
+      "learning_rate": 0.0002,
+      "loss": 0.9022,
+      "step": 11220
+    },
+    {
+      "epoch": 0.7369975389663659,
+      "grad_norm": 0.412392795085907,
+      "learning_rate": 0.0002,
+      "loss": 0.9171,
+      "step": 11230
+    },
+    {
+      "epoch": 0.7376538146021329,
+      "grad_norm": 0.4415507912635803,
+      "learning_rate": 0.0002,
+      "loss": 0.8736,
+      "step": 11240
+    },
+    {
+      "epoch": 0.7383100902379,
+      "grad_norm": 0.35013696551322937,
+      "learning_rate": 0.0002,
+      "loss": 0.8936,
+      "step": 11250
+    },
+    {
+      "epoch": 0.7389663658736669,
+      "grad_norm": 0.3677300810813904,
+      "learning_rate": 0.0002,
+      "loss": 0.8841,
+      "step": 11260
+    },
+    {
+      "epoch": 0.7396226415094339,
+      "grad_norm": 0.36722511053085327,
+      "learning_rate": 0.0002,
+      "loss": 0.8756,
+      "step": 11270
+    },
+    {
+      "epoch": 0.740278917145201,
+      "grad_norm": 0.41611534357070923,
+      "learning_rate": 0.0002,
+      "loss": 0.8727,
+      "step": 11280
+    },
+    {
+      "epoch": 0.740935192780968,
+      "grad_norm": 0.4244968295097351,
+      "learning_rate": 0.0002,
+      "loss": 0.921,
+      "step": 11290
+    },
+    {
+      "epoch": 0.7415914684167351,
+      "grad_norm": 0.38986068964004517,
+      "learning_rate": 0.0002,
+      "loss": 0.9696,
+      "step": 11300
+    },
+    {
+      "epoch": 0.742247744052502,
+      "grad_norm": 0.33212459087371826,
+      "learning_rate": 0.0002,
+      "loss": 0.855,
+      "step": 11310
+    },
+    {
+      "epoch": 0.742904019688269,
+      "grad_norm": 0.4290331304073334,
+      "learning_rate": 0.0002,
+      "loss": 0.9296,
+      "step": 11320
+    },
+    {
+      "epoch": 0.7435602953240361,
+      "grad_norm": 0.3750900626182556,
+      "learning_rate": 0.0002,
+      "loss": 0.867,
+      "step": 11330
+    },
+    {
+      "epoch": 0.7442165709598031,
+      "grad_norm": 0.37758177518844604,
+      "learning_rate": 0.0002,
+      "loss": 0.9462,
+      "step": 11340
+    },
+    {
+      "epoch": 0.7448728465955702,
+      "grad_norm": 0.31147271394729614,
+      "learning_rate": 0.0002,
+      "loss": 0.849,
+      "step": 11350
+    },
+    {
+      "epoch": 0.7455291222313372,
+      "grad_norm": 0.4493428170681,
+      "learning_rate": 0.0002,
+      "loss": 0.9675,
+      "step": 11360
+    },
+    {
+      "epoch": 0.7461853978671041,
+      "grad_norm": 0.4268129765987396,
+      "learning_rate": 0.0002,
+      "loss": 0.9506,
+      "step": 11370
+    },
+    {
+      "epoch": 0.7468416735028712,
+      "grad_norm": 0.3716314733028412,
+      "learning_rate": 0.0002,
+      "loss": 0.9371,
+      "step": 11380
+    },
+    {
+      "epoch": 0.7474979491386382,
+      "grad_norm": 0.33728593587875366,
+      "learning_rate": 0.0002,
+      "loss": 0.8762,
+      "step": 11390
+    },
+    {
+      "epoch": 0.7481542247744053,
+      "grad_norm": 0.36548155546188354,
+      "learning_rate": 0.0002,
+      "loss": 0.8532,
+      "step": 11400
+    },
+    {
+      "epoch": 0.7488105004101723,
+      "grad_norm": 0.32645145058631897,
+      "learning_rate": 0.0002,
+      "loss": 0.8709,
+      "step": 11410
+    },
+    {
+      "epoch": 0.7494667760459393,
+      "grad_norm": 0.36403900384902954,
+      "learning_rate": 0.0002,
+      "loss": 0.9169,
+      "step": 11420
+    },
+    {
+      "epoch": 0.7501230516817063,
+      "grad_norm": 0.4186992049217224,
+      "learning_rate": 0.0002,
+      "loss": 0.9734,
+      "step": 11430
+    },
+    {
+      "epoch": 0.7507793273174733,
+      "grad_norm": 0.342457115650177,
+      "learning_rate": 0.0002,
+      "loss": 0.8982,
+      "step": 11440
+    },
+    {
+      "epoch": 0.7514356029532404,
+      "grad_norm": 0.3882320821285248,
+      "learning_rate": 0.0002,
+      "loss": 0.8785,
+      "step": 11450
+    },
+    {
+      "epoch": 0.7520918785890074,
+      "grad_norm": 0.4653763771057129,
+      "learning_rate": 0.0002,
+      "loss": 0.9492,
+      "step": 11460
+    },
+    {
+      "epoch": 0.7527481542247744,
+      "grad_norm": 0.3960241377353668,
+      "learning_rate": 0.0002,
+      "loss": 0.9357,
+      "step": 11470
+    },
+    {
+      "epoch": 0.7534044298605415,
+      "grad_norm": 0.38367652893066406,
+      "learning_rate": 0.0002,
+      "loss": 0.8912,
+      "step": 11480
+    },
+    {
+      "epoch": 0.7540607054963084,
+      "grad_norm": 0.39881640672683716,
+      "learning_rate": 0.0002,
+      "loss": 0.9367,
+      "step": 11490
+    },
+    {
+      "epoch": 0.7547169811320755,
+      "grad_norm": 0.40201085805892944,
+      "learning_rate": 0.0002,
+      "loss": 0.9018,
+      "step": 11500
+    },
+    {
+      "epoch": 0.7553732567678425,
+      "grad_norm": 0.38070711493492126,
+      "learning_rate": 0.0002,
+      "loss": 0.8668,
+      "step": 11510
+    },
+    {
+      "epoch": 0.7560295324036095,
+      "grad_norm": 0.334245502948761,
+      "learning_rate": 0.0002,
+      "loss": 0.8542,
+      "step": 11520
+    },
+    {
+      "epoch": 0.7566858080393766,
+      "grad_norm": 0.42725566029548645,
+      "learning_rate": 0.0002,
+      "loss": 0.9036,
+      "step": 11530
+    },
+    {
+      "epoch": 0.7573420836751436,
+      "grad_norm": 1.4140206575393677,
+      "learning_rate": 0.0002,
+      "loss": 0.9443,
+      "step": 11540
+    },
+    {
+      "epoch": 0.7579983593109105,
+      "grad_norm": 0.5802565813064575,
+      "learning_rate": 0.0002,
+      "loss": 0.8962,
+      "step": 11550
+    },
+    {
+      "epoch": 0.7586546349466776,
+      "grad_norm": 0.45892661809921265,
+      "learning_rate": 0.0002,
+      "loss": 0.9095,
+      "step": 11560
+    },
+    {
+      "epoch": 0.7593109105824446,
+      "grad_norm": 0.3837396800518036,
+      "learning_rate": 0.0002,
+      "loss": 0.8373,
+      "step": 11570
+    },
+    {
+      "epoch": 0.7599671862182117,
+      "grad_norm": 0.4158342480659485,
+      "learning_rate": 0.0002,
+      "loss": 0.9452,
+      "step": 11580
+    },
+    {
+      "epoch": 0.7606234618539787,
+      "grad_norm": 0.4374973773956299,
+      "learning_rate": 0.0002,
+      "loss": 0.9551,
+      "step": 11590
+    },
+    {
+      "epoch": 0.7612797374897456,
+      "grad_norm": 0.36342450976371765,
+      "learning_rate": 0.0002,
+      "loss": 0.9181,
+      "step": 11600
+    },
+    {
+      "epoch": 0.7619360131255127,
+      "grad_norm": 0.37019461393356323,
+      "learning_rate": 0.0002,
+      "loss": 0.8596,
+      "step": 11610
+    },
+    {
+      "epoch": 0.7625922887612797,
+      "grad_norm": 0.3745611608028412,
+      "learning_rate": 0.0002,
+      "loss": 0.8804,
+      "step": 11620
+    },
+    {
+      "epoch": 0.7632485643970468,
+      "grad_norm": 0.33488842844963074,
+      "learning_rate": 0.0002,
+      "loss": 0.8745,
+      "step": 11630
+    },
+    {
+      "epoch": 0.7639048400328138,
+      "grad_norm": 0.3700532615184784,
+      "learning_rate": 0.0002,
+      "loss": 0.8798,
+      "step": 11640
+    },
+    {
+      "epoch": 0.7645611156685808,
+      "grad_norm": 0.3722131848335266,
+      "learning_rate": 0.0002,
+      "loss": 0.8943,
+      "step": 11650
+    },
+    {
+      "epoch": 0.7652173913043478,
+      "grad_norm": 0.3463144302368164,
+      "learning_rate": 0.0002,
+      "loss": 0.8702,
+      "step": 11660
+    },
+    {
+      "epoch": 0.7658736669401148,
+      "grad_norm": 0.339691162109375,
+      "learning_rate": 0.0002,
+      "loss": 0.865,
+      "step": 11670
+    },
+    {
+      "epoch": 0.7665299425758819,
+      "grad_norm": 0.33323508501052856,
+      "learning_rate": 0.0002,
+      "loss": 0.9536,
+      "step": 11680
+    },
+    {
+      "epoch": 0.7671862182116489,
+      "grad_norm": 0.3937166631221771,
+      "learning_rate": 0.0002,
+      "loss": 0.8657,
+      "step": 11690
+    },
+    {
+      "epoch": 0.7678424938474159,
+      "grad_norm": 0.4112081527709961,
+      "learning_rate": 0.0002,
+      "loss": 0.8857,
+      "step": 11700
+    },
+    {
+      "epoch": 0.768498769483183,
+      "grad_norm": 0.4242405593395233,
+      "learning_rate": 0.0002,
+      "loss": 0.9108,
+      "step": 11710
+    },
+    {
+      "epoch": 0.7691550451189499,
+      "grad_norm": 0.33512821793556213,
+      "learning_rate": 0.0002,
+      "loss": 0.9148,
+      "step": 11720
+    },
+    {
+      "epoch": 0.769811320754717,
+      "grad_norm": 0.5148407816886902,
+      "learning_rate": 0.0002,
+      "loss": 0.8599,
+      "step": 11730
+    },
+    {
+      "epoch": 0.770467596390484,
+      "grad_norm": 0.4562109112739563,
+      "learning_rate": 0.0002,
+      "loss": 0.9678,
+      "step": 11740
+    },
+    {
+      "epoch": 0.771123872026251,
+      "grad_norm": 0.4246354401111603,
+      "learning_rate": 0.0002,
+      "loss": 0.8937,
+      "step": 11750
+    },
+    {
+      "epoch": 0.7717801476620181,
+      "grad_norm": 0.39338991045951843,
+      "learning_rate": 0.0002,
+      "loss": 0.8821,
+      "step": 11760
+    },
+    {
+      "epoch": 0.7724364232977851,
+      "grad_norm": 0.403199166059494,
+      "learning_rate": 0.0002,
+      "loss": 0.8769,
+      "step": 11770
+    },
+    {
+      "epoch": 0.7730926989335521,
+      "grad_norm": 0.4494798183441162,
+      "learning_rate": 0.0002,
+      "loss": 0.8914,
+      "step": 11780
+    },
+    {
+      "epoch": 0.7737489745693191,
+      "grad_norm": 0.3649079501628876,
+      "learning_rate": 0.0002,
+      "loss": 0.8633,
+      "step": 11790
+    },
+    {
+      "epoch": 0.7744052502050861,
+      "grad_norm": 0.3314788341522217,
+      "learning_rate": 0.0002,
+      "loss": 0.8684,
+      "step": 11800
+    },
+    {
+      "epoch": 0.7750615258408532,
+      "grad_norm": 0.36505308747291565,
+      "learning_rate": 0.0002,
+      "loss": 0.8756,
+      "step": 11810
+    },
+    {
+      "epoch": 0.7757178014766202,
+      "grad_norm": 0.45973560214042664,
+      "learning_rate": 0.0002,
+      "loss": 0.8778,
+      "step": 11820
+    },
+    {
+      "epoch": 0.7763740771123872,
+      "grad_norm": 0.3941294550895691,
+      "learning_rate": 0.0002,
+      "loss": 0.8845,
+      "step": 11830
+    },
+    {
+      "epoch": 0.7770303527481542,
+      "grad_norm": 0.34385251998901367,
+      "learning_rate": 0.0002,
+      "loss": 0.9033,
+      "step": 11840
+    },
+    {
+      "epoch": 0.7776866283839212,
+      "grad_norm": 0.35967403650283813,
+      "learning_rate": 0.0002,
+      "loss": 0.9595,
+      "step": 11850
+    },
+    {
+      "epoch": 0.7783429040196883,
+      "grad_norm": 0.4731179177761078,
+      "learning_rate": 0.0002,
+      "loss": 0.9193,
+      "step": 11860
+    },
+    {
+      "epoch": 0.7789991796554553,
+      "grad_norm": 0.38854387402534485,
+      "learning_rate": 0.0002,
+      "loss": 0.9344,
+      "step": 11870
+    },
+    {
+      "epoch": 0.7796554552912223,
+      "grad_norm": 0.3925110995769501,
+      "learning_rate": 0.0002,
+      "loss": 0.9056,
+      "step": 11880
+    },
+    {
+      "epoch": 0.7803117309269894,
+      "grad_norm": 0.35882773995399475,
+      "learning_rate": 0.0002,
+      "loss": 0.8818,
+      "step": 11890
+    },
+    {
+      "epoch": 0.7809680065627563,
+      "grad_norm": 0.4351222813129425,
+      "learning_rate": 0.0002,
+      "loss": 0.9097,
+      "step": 11900
+    },
+    {
+      "epoch": 0.7816242821985234,
+      "grad_norm": 0.39528653025627136,
+      "learning_rate": 0.0002,
+      "loss": 0.8866,
+      "step": 11910
+    },
+    {
+      "epoch": 0.7822805578342904,
+      "grad_norm": 0.34876471757888794,
+      "learning_rate": 0.0002,
+      "loss": 0.91,
+      "step": 11920
+    },
+    {
+      "epoch": 0.7829368334700574,
+      "grad_norm": 0.44766634702682495,
+      "learning_rate": 0.0002,
+      "loss": 0.9329,
+      "step": 11930
+    },
+    {
+      "epoch": 0.7835931091058245,
+      "grad_norm": 0.42268314957618713,
+      "learning_rate": 0.0002,
+      "loss": 0.9687,
+      "step": 11940
+    },
+    {
+      "epoch": 0.7842493847415914,
+      "grad_norm": 0.377101868391037,
+      "learning_rate": 0.0002,
+      "loss": 0.911,
+      "step": 11950
+    },
+    {
+      "epoch": 0.7849056603773585,
+      "grad_norm": 0.4489518404006958,
+      "learning_rate": 0.0002,
+      "loss": 0.8711,
+      "step": 11960
+    },
+    {
+      "epoch": 0.7855619360131255,
+      "grad_norm": 0.4585464596748352,
+      "learning_rate": 0.0002,
+      "loss": 0.9177,
+      "step": 11970
+    },
+    {
+      "epoch": 0.7862182116488925,
+      "grad_norm": 0.41329675912857056,
+      "learning_rate": 0.0002,
+      "loss": 0.901,
+      "step": 11980
+    },
+    {
+      "epoch": 0.7868744872846596,
+      "grad_norm": 0.42738014459609985,
+      "learning_rate": 0.0002,
+      "loss": 0.8993,
+      "step": 11990
+    },
+    {
+      "epoch": 0.7875307629204266,
+      "grad_norm": 0.40386950969696045,
+      "learning_rate": 0.0002,
+      "loss": 0.9206,
+      "step": 12000
+    },
+    {
+      "epoch": 0.7881870385561937,
+      "grad_norm": 0.3859177827835083,
+      "learning_rate": 0.0002,
+      "loss": 0.8702,
+      "step": 12010
+    },
+    {
+      "epoch": 0.7888433141919606,
+      "grad_norm": 0.3537571430206299,
+      "learning_rate": 0.0002,
+      "loss": 0.87,
+      "step": 12020
+    },
+    {
+      "epoch": 0.7894995898277276,
+      "grad_norm": 0.39319470524787903,
+      "learning_rate": 0.0002,
+      "loss": 0.9106,
+      "step": 12030
+    },
+    {
+      "epoch": 0.7901558654634947,
+      "grad_norm": 0.32762688398361206,
+      "learning_rate": 0.0002,
+      "loss": 0.9197,
+      "step": 12040
+    },
+    {
+      "epoch": 0.7908121410992617,
+      "grad_norm": 0.34617939591407776,
+      "learning_rate": 0.0002,
+      "loss": 0.9218,
+      "step": 12050
+    },
+    {
+      "epoch": 0.7914684167350287,
+      "grad_norm": 0.4012453258037567,
+      "learning_rate": 0.0002,
+      "loss": 0.9099,
+      "step": 12060
+    },
+    {
+      "epoch": 0.7921246923707957,
+      "grad_norm": 0.4265747666358948,
+      "learning_rate": 0.0002,
+      "loss": 0.9857,
+      "step": 12070
+    },
+    {
+      "epoch": 0.7927809680065627,
+      "grad_norm": 0.42164385318756104,
+      "learning_rate": 0.0002,
+      "loss": 0.9169,
+      "step": 12080
+    },
+    {
+      "epoch": 0.7934372436423298,
+      "grad_norm": 0.4054335951805115,
+      "learning_rate": 0.0002,
+      "loss": 0.8952,
+      "step": 12090
+    },
+    {
+      "epoch": 0.7940935192780968,
+      "grad_norm": 0.38486307859420776,
+      "learning_rate": 0.0002,
+      "loss": 0.9706,
+      "step": 12100
+    },
+    {
+      "epoch": 0.7947497949138638,
+      "grad_norm": 0.3842015266418457,
+      "learning_rate": 0.0002,
+      "loss": 0.8703,
+      "step": 12110
+    },
+    {
+      "epoch": 0.7954060705496309,
+      "grad_norm": 0.3772023320198059,
+      "learning_rate": 0.0002,
+      "loss": 0.8847,
+      "step": 12120
+    },
+    {
+      "epoch": 0.7960623461853978,
+      "grad_norm": 0.39477309584617615,
+      "learning_rate": 0.0002,
+      "loss": 0.9248,
+      "step": 12130
+    },
+    {
+      "epoch": 0.7967186218211649,
+      "grad_norm": 0.3458614945411682,
+      "learning_rate": 0.0002,
+      "loss": 0.8865,
+      "step": 12140
+    },
+    {
+      "epoch": 0.7973748974569319,
+      "grad_norm": 0.42238670587539673,
+      "learning_rate": 0.0002,
+      "loss": 0.884,
+      "step": 12150
+    },
+    {
+      "epoch": 0.7980311730926989,
+      "grad_norm": 0.3623220920562744,
+      "learning_rate": 0.0002,
+      "loss": 0.9561,
+      "step": 12160
+    },
+    {
+      "epoch": 0.798687448728466,
+      "grad_norm": 0.426715224981308,
+      "learning_rate": 0.0002,
+      "loss": 1.0033,
+      "step": 12170
+    },
+    {
+      "epoch": 0.799343724364233,
+      "grad_norm": 0.3558938205242157,
+      "learning_rate": 0.0002,
+      "loss": 0.9158,
+      "step": 12180
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.426761269569397,
+      "learning_rate": 0.0002,
+      "loss": 0.9323,
+      "step": 12190
+    },
+    {
+      "epoch": 0.800656275635767,
+      "grad_norm": 0.42333319783210754,
+      "learning_rate": 0.0002,
+      "loss": 0.9256,
+      "step": 12200
+    },
+    {
+      "epoch": 0.801312551271534,
+      "grad_norm": 0.42534688115119934,
+      "learning_rate": 0.0002,
+      "loss": 0.9307,
+      "step": 12210
+    },
+    {
+      "epoch": 0.8019688269073011,
+      "grad_norm": 0.37565773725509644,
+      "learning_rate": 0.0002,
+      "loss": 0.8999,
+      "step": 12220
+    },
+    {
+      "epoch": 0.8026251025430681,
+      "grad_norm": 0.3591224253177643,
+      "learning_rate": 0.0002,
+      "loss": 0.8531,
+      "step": 12230
+    },
+    {
+      "epoch": 0.8032813781788352,
+      "grad_norm": 0.35887411236763,
+      "learning_rate": 0.0002,
+      "loss": 0.8817,
+      "step": 12240
+    },
+    {
+      "epoch": 0.8039376538146021,
+      "grad_norm": 0.3895672559738159,
+      "learning_rate": 0.0002,
+      "loss": 0.8562,
+      "step": 12250
+    },
+    {
+      "epoch": 0.8045939294503691,
+      "grad_norm": 0.3483835756778717,
+      "learning_rate": 0.0002,
+      "loss": 0.8885,
+      "step": 12260
+    },
+    {
+      "epoch": 0.8052502050861362,
+      "grad_norm": 0.37694090604782104,
+      "learning_rate": 0.0002,
+      "loss": 0.9343,
+      "step": 12270
+    },
+    {
+      "epoch": 0.8059064807219032,
+      "grad_norm": 0.4011424779891968,
+      "learning_rate": 0.0002,
+      "loss": 0.9315,
+      "step": 12280
+    },
+    {
+      "epoch": 0.8065627563576703,
+      "grad_norm": 0.3254278004169464,
+      "learning_rate": 0.0002,
+      "loss": 0.9341,
+      "step": 12290
+    },
+    {
+      "epoch": 0.8072190319934373,
+      "grad_norm": 0.3868531286716461,
+      "learning_rate": 0.0002,
+      "loss": 0.9025,
+      "step": 12300
+    },
+    {
+      "epoch": 0.8078753076292042,
+      "grad_norm": 0.44830775260925293,
+      "learning_rate": 0.0002,
+      "loss": 0.8959,
+      "step": 12310
+    },
+    {
+      "epoch": 0.8085315832649713,
+      "grad_norm": 0.38808006048202515,
+      "learning_rate": 0.0002,
+      "loss": 0.9639,
+      "step": 12320
+    },
+    {
+      "epoch": 0.8091878589007383,
+      "grad_norm": 0.42684856057167053,
+      "learning_rate": 0.0002,
+      "loss": 0.9119,
+      "step": 12330
+    },
+    {
+      "epoch": 0.8098441345365053,
+      "grad_norm": 0.310099333524704,
+      "learning_rate": 0.0002,
+      "loss": 0.9231,
+      "step": 12340
+    },
+    {
+      "epoch": 0.8105004101722724,
+      "grad_norm": 0.5252287983894348,
+      "learning_rate": 0.0002,
+      "loss": 0.8607,
+      "step": 12350
+    },
+    {
+      "epoch": 0.8111566858080393,
+      "grad_norm": 0.41571253538131714,
+      "learning_rate": 0.0002,
+      "loss": 0.8559,
+      "step": 12360
+    },
+    {
+      "epoch": 0.8118129614438064,
+      "grad_norm": 0.4471582770347595,
+      "learning_rate": 0.0002,
+      "loss": 0.8626,
+      "step": 12370
+    },
+    {
+      "epoch": 0.8124692370795734,
+      "grad_norm": 0.3977353274822235,
+      "learning_rate": 0.0002,
+      "loss": 0.923,
+      "step": 12380
+    },
+    {
+      "epoch": 0.8131255127153404,
+      "grad_norm": 0.3274862766265869,
+      "learning_rate": 0.0002,
+      "loss": 0.8792,
+      "step": 12390
+    },
+    {
+      "epoch": 0.8137817883511075,
+      "grad_norm": 0.4114132821559906,
+      "learning_rate": 0.0002,
+      "loss": 0.907,
+      "step": 12400
+    },
+    {
+      "epoch": 0.8144380639868745,
+      "grad_norm": 0.40929168462753296,
+      "learning_rate": 0.0002,
+      "loss": 0.9324,
+      "step": 12410
+    },
+    {
+      "epoch": 0.8150943396226416,
+      "grad_norm": 0.32346615195274353,
+      "learning_rate": 0.0002,
+      "loss": 0.8745,
+      "step": 12420
+    },
+    {
+      "epoch": 0.8157506152584085,
+      "grad_norm": 0.36344656348228455,
+      "learning_rate": 0.0002,
+      "loss": 0.8766,
+      "step": 12430
+    },
+    {
+      "epoch": 0.8164068908941755,
+      "grad_norm": 0.3749464452266693,
+      "learning_rate": 0.0002,
+      "loss": 0.872,
+      "step": 12440
+    },
+    {
+      "epoch": 0.8170631665299426,
+      "grad_norm": 0.4742373526096344,
+      "learning_rate": 0.0002,
+      "loss": 0.9128,
+      "step": 12450
+    },
+    {
+      "epoch": 0.8177194421657096,
+      "grad_norm": 0.3863218128681183,
+      "learning_rate": 0.0002,
+      "loss": 0.8972,
+      "step": 12460
+    },
+    {
+      "epoch": 0.8183757178014767,
+      "grad_norm": 0.47756487131118774,
+      "learning_rate": 0.0002,
+      "loss": 0.9148,
+      "step": 12470
+    },
+    {
+      "epoch": 0.8190319934372436,
+      "grad_norm": 0.4045886993408203,
+      "learning_rate": 0.0002,
+      "loss": 0.8817,
+      "step": 12480
+    },
+    {
+      "epoch": 0.8196882690730106,
+      "grad_norm": 0.4119892418384552,
+      "learning_rate": 0.0002,
+      "loss": 0.866,
+      "step": 12490
+    },
+    {
+      "epoch": 0.8203445447087777,
+      "grad_norm": 0.3714514672756195,
+      "learning_rate": 0.0002,
+      "loss": 0.866,
+      "step": 12500
+    },
+    {
+      "epoch": 0.8210008203445447,
+      "grad_norm": 0.38410791754722595,
+      "learning_rate": 0.0002,
+      "loss": 0.8943,
+      "step": 12510
+    },
+    {
+      "epoch": 0.8216570959803118,
+      "grad_norm": 0.3178478479385376,
+      "learning_rate": 0.0002,
+      "loss": 0.8857,
+      "step": 12520
+    },
+    {
+      "epoch": 0.8223133716160788,
+      "grad_norm": 0.4347972869873047,
+      "learning_rate": 0.0002,
+      "loss": 0.9356,
+      "step": 12530
+    },
+    {
+      "epoch": 0.8229696472518457,
+      "grad_norm": 0.4276008903980255,
+      "learning_rate": 0.0002,
+      "loss": 0.8926,
+      "step": 12540
+    },
+    {
+      "epoch": 0.8236259228876128,
+      "grad_norm": 0.36309465765953064,
+      "learning_rate": 0.0002,
+      "loss": 0.9213,
+      "step": 12550
+    },
+    {
+      "epoch": 0.8242821985233798,
+      "grad_norm": 0.45721492171287537,
+      "learning_rate": 0.0002,
+      "loss": 0.8847,
+      "step": 12560
+    },
+    {
+      "epoch": 0.8249384741591469,
+      "grad_norm": 0.37675052881240845,
+      "learning_rate": 0.0002,
+      "loss": 0.8655,
+      "step": 12570
+    },
+    {
+      "epoch": 0.8255947497949139,
+      "grad_norm": 0.41907957196235657,
+      "learning_rate": 0.0002,
+      "loss": 0.8717,
+      "step": 12580
+    },
+    {
+      "epoch": 0.8262510254306809,
+      "grad_norm": 0.36430326104164124,
+      "learning_rate": 0.0002,
+      "loss": 0.9465,
+      "step": 12590
+    },
+    {
+      "epoch": 0.8269073010664479,
+      "grad_norm": 0.45015767216682434,
+      "learning_rate": 0.0002,
+      "loss": 0.9244,
+      "step": 12600
+    },
+    {
+      "epoch": 0.8275635767022149,
+      "grad_norm": 0.369115948677063,
+      "learning_rate": 0.0002,
+      "loss": 0.9076,
+      "step": 12610
+    },
+    {
+      "epoch": 0.8282198523379819,
+      "grad_norm": 0.4821915030479431,
+      "learning_rate": 0.0002,
+      "loss": 0.92,
+      "step": 12620
+    },
+    {
+      "epoch": 0.828876127973749,
+      "grad_norm": 0.4291541874408722,
+      "learning_rate": 0.0002,
+      "loss": 0.8813,
+      "step": 12630
+    },
+    {
+      "epoch": 0.829532403609516,
+      "grad_norm": 0.3816904127597809,
+      "learning_rate": 0.0002,
+      "loss": 0.9196,
+      "step": 12640
+    },
+    {
+      "epoch": 0.8301886792452831,
+      "grad_norm": 0.3915407657623291,
+      "learning_rate": 0.0002,
+      "loss": 0.8821,
+      "step": 12650
+    },
+    {
+      "epoch": 0.83084495488105,
+      "grad_norm": 0.35964086651802063,
+      "learning_rate": 0.0002,
+      "loss": 0.907,
+      "step": 12660
+    },
+    {
+      "epoch": 0.831501230516817,
+      "grad_norm": 0.3201609253883362,
+      "learning_rate": 0.0002,
+      "loss": 0.8512,
+      "step": 12670
+    },
+    {
+      "epoch": 0.8321575061525841,
+      "grad_norm": 0.44153767824172974,
+      "learning_rate": 0.0002,
+      "loss": 0.9085,
+      "step": 12680
+    },
+    {
+      "epoch": 0.8328137817883511,
+      "grad_norm": 0.37000054121017456,
+      "learning_rate": 0.0002,
+      "loss": 0.859,
+      "step": 12690
+    },
+    {
+      "epoch": 0.8334700574241182,
+      "grad_norm": 0.39843010902404785,
+      "learning_rate": 0.0002,
+      "loss": 0.9413,
+      "step": 12700
+    },
+    {
+      "epoch": 0.8341263330598851,
+      "grad_norm": 0.3594053387641907,
+      "learning_rate": 0.0002,
+      "loss": 0.8787,
+      "step": 12710
+    },
+    {
+      "epoch": 0.8347826086956521,
+      "grad_norm": 0.38574180006980896,
+      "learning_rate": 0.0002,
+      "loss": 0.8542,
+      "step": 12720
+    },
+    {
+      "epoch": 0.8354388843314192,
+      "grad_norm": 0.4101716876029968,
+      "learning_rate": 0.0002,
+      "loss": 0.8746,
+      "step": 12730
+    },
+    {
+      "epoch": 0.8360951599671862,
+      "grad_norm": 0.3665215075016022,
+      "learning_rate": 0.0002,
+      "loss": 0.8442,
+      "step": 12740
+    },
+    {
+      "epoch": 0.8367514356029533,
+      "grad_norm": 0.39471596479415894,
+      "learning_rate": 0.0002,
+      "loss": 0.8898,
+      "step": 12750
+    },
+    {
+      "epoch": 0.8374077112387203,
+      "grad_norm": 0.35624340176582336,
+      "learning_rate": 0.0002,
+      "loss": 0.9026,
+      "step": 12760
+    },
+    {
+      "epoch": 0.8380639868744872,
+      "grad_norm": 0.3838249742984772,
+      "learning_rate": 0.0002,
+      "loss": 0.917,
+      "step": 12770
+    },
+    {
+      "epoch": 0.8387202625102543,
+      "grad_norm": 0.408368319272995,
+      "learning_rate": 0.0002,
+      "loss": 0.874,
+      "step": 12780
+    },
+    {
+      "epoch": 0.8393765381460213,
+      "grad_norm": 0.46758291125297546,
+      "learning_rate": 0.0002,
+      "loss": 0.934,
+      "step": 12790
+    },
+    {
+      "epoch": 0.8400328137817884,
+      "grad_norm": 0.35787731409072876,
+      "learning_rate": 0.0002,
+      "loss": 0.8603,
+      "step": 12800
+    },
+    {
+      "epoch": 0.8406890894175554,
+      "grad_norm": 0.39618661999702454,
+      "learning_rate": 0.0002,
+      "loss": 0.9026,
+      "step": 12810
+    },
+    {
+      "epoch": 0.8413453650533224,
+      "grad_norm": 0.44962066411972046,
+      "learning_rate": 0.0002,
+      "loss": 0.9307,
+      "step": 12820
+    },
+    {
+      "epoch": 0.8420016406890894,
+      "grad_norm": 0.36435529589653015,
+      "learning_rate": 0.0002,
+      "loss": 0.8756,
+      "step": 12830
+    },
+    {
+      "epoch": 0.8426579163248564,
+      "grad_norm": 0.37484753131866455,
+      "learning_rate": 0.0002,
+      "loss": 0.93,
+      "step": 12840
+    },
+    {
+      "epoch": 0.8433141919606235,
+      "grad_norm": 0.36679843068122864,
+      "learning_rate": 0.0002,
+      "loss": 0.8655,
+      "step": 12850
+    },
+    {
+      "epoch": 0.8439704675963905,
+      "grad_norm": 0.3948156237602234,
+      "learning_rate": 0.0002,
+      "loss": 0.9339,
+      "step": 12860
+    },
+    {
+      "epoch": 0.8446267432321575,
+      "grad_norm": 0.3789501488208771,
+      "learning_rate": 0.0002,
+      "loss": 0.8915,
+      "step": 12870
+    },
+    {
+      "epoch": 0.8452830188679246,
+      "grad_norm": 0.45156800746917725,
+      "learning_rate": 0.0002,
+      "loss": 0.916,
+      "step": 12880
+    },
+    {
+      "epoch": 0.8459392945036915,
+      "grad_norm": 0.37762370705604553,
+      "learning_rate": 0.0002,
+      "loss": 0.8762,
+      "step": 12890
+    },
+    {
+      "epoch": 0.8465955701394585,
+      "grad_norm": 0.41654065251350403,
+      "learning_rate": 0.0002,
+      "loss": 0.8922,
+      "step": 12900
+    },
+    {
+      "epoch": 0.8472518457752256,
+      "grad_norm": 0.35026174783706665,
+      "learning_rate": 0.0002,
+      "loss": 0.9136,
+      "step": 12910
+    },
+    {
+      "epoch": 0.8479081214109926,
+      "grad_norm": 0.5227314829826355,
+      "learning_rate": 0.0002,
+      "loss": 0.8443,
+      "step": 12920
+    },
+    {
+      "epoch": 0.8485643970467597,
+      "grad_norm": 0.365063339471817,
+      "learning_rate": 0.0002,
+      "loss": 0.8533,
+      "step": 12930
+    },
+    {
+      "epoch": 0.8492206726825267,
+      "grad_norm": 0.40983277559280396,
+      "learning_rate": 0.0002,
+      "loss": 0.9012,
+      "step": 12940
+    },
+    {
+      "epoch": 0.8498769483182936,
+      "grad_norm": 0.36414071917533875,
+      "learning_rate": 0.0002,
+      "loss": 0.8788,
+      "step": 12950
+    },
+    {
+      "epoch": 0.8505332239540607,
+      "grad_norm": 0.3703882694244385,
+      "learning_rate": 0.0002,
+      "loss": 0.8981,
+      "step": 12960
+    },
+    {
+      "epoch": 0.8511894995898277,
+      "grad_norm": 0.3218643367290497,
+      "learning_rate": 0.0002,
+      "loss": 0.9256,
+      "step": 12970
+    },
+    {
+      "epoch": 0.8518457752255948,
+      "grad_norm": 0.38008660078048706,
+      "learning_rate": 0.0002,
+      "loss": 0.8525,
+      "step": 12980
+    },
+    {
+      "epoch": 0.8525020508613618,
+      "grad_norm": 0.4357127547264099,
+      "learning_rate": 0.0002,
+      "loss": 0.8651,
+      "step": 12990
+    },
+    {
+      "epoch": 0.8531583264971287,
+      "grad_norm": 0.36831775307655334,
+      "learning_rate": 0.0002,
+      "loss": 0.8592,
+      "step": 13000
+    },
+    {
+      "epoch": 0.8538146021328958,
+      "grad_norm": 0.4546806812286377,
+      "learning_rate": 0.0002,
+      "loss": 0.8772,
+      "step": 13010
+    },
+    {
+      "epoch": 0.8544708777686628,
+      "grad_norm": 0.3474937379360199,
+      "learning_rate": 0.0002,
+      "loss": 0.865,
+      "step": 13020
+    },
+    {
+      "epoch": 0.8551271534044299,
+      "grad_norm": 0.438834547996521,
+      "learning_rate": 0.0002,
+      "loss": 0.9098,
+      "step": 13030
+    },
+    {
+      "epoch": 0.8557834290401969,
+      "grad_norm": 0.3401171565055847,
+      "learning_rate": 0.0002,
+      "loss": 0.9573,
+      "step": 13040
+    },
+    {
+      "epoch": 0.8564397046759639,
+      "grad_norm": 0.3887326419353485,
+      "learning_rate": 0.0002,
+      "loss": 0.8286,
+      "step": 13050
+    },
+    {
+      "epoch": 0.857095980311731,
+      "grad_norm": 0.3489287197589874,
+      "learning_rate": 0.0002,
+      "loss": 0.8695,
+      "step": 13060
+    },
+    {
+      "epoch": 0.8577522559474979,
+      "grad_norm": 0.481189489364624,
+      "learning_rate": 0.0002,
+      "loss": 0.9118,
+      "step": 13070
+    },
+    {
+      "epoch": 0.858408531583265,
+      "grad_norm": 0.4638312757015228,
+      "learning_rate": 0.0002,
+      "loss": 0.9116,
+      "step": 13080
+    },
+    {
+      "epoch": 0.859064807219032,
+      "grad_norm": 0.43477529287338257,
+      "learning_rate": 0.0002,
+      "loss": 0.8775,
+      "step": 13090
+    },
+    {
+      "epoch": 0.859721082854799,
+      "grad_norm": 0.43358466029167175,
+      "learning_rate": 0.0002,
+      "loss": 0.9232,
+      "step": 13100
+    },
+    {
+      "epoch": 0.8603773584905661,
+      "grad_norm": 0.385527104139328,
+      "learning_rate": 0.0002,
+      "loss": 0.8771,
+      "step": 13110
+    },
+    {
+      "epoch": 0.861033634126333,
+      "grad_norm": 0.37878429889678955,
+      "learning_rate": 0.0002,
+      "loss": 0.8708,
+      "step": 13120
+    },
+    {
+      "epoch": 0.8616899097621001,
+      "grad_norm": 0.409476637840271,
+      "learning_rate": 0.0002,
+      "loss": 0.83,
+      "step": 13130
+    },
+    {
+      "epoch": 0.8623461853978671,
+      "grad_norm": 0.3888716697692871,
+      "learning_rate": 0.0002,
+      "loss": 0.9149,
+      "step": 13140
+    },
+    {
+      "epoch": 0.8630024610336341,
+      "grad_norm": 0.38093528151512146,
+      "learning_rate": 0.0002,
+      "loss": 0.8261,
+      "step": 13150
+    },
+    {
+      "epoch": 0.8636587366694012,
+      "grad_norm": 0.45613282918930054,
+      "learning_rate": 0.0002,
+      "loss": 0.9163,
+      "step": 13160
+    },
+    {
+      "epoch": 0.8643150123051682,
+      "grad_norm": 0.41665518283843994,
+      "learning_rate": 0.0002,
+      "loss": 0.8967,
+      "step": 13170
+    },
+    {
+      "epoch": 0.8649712879409351,
+      "grad_norm": 0.4322538673877716,
+      "learning_rate": 0.0002,
+      "loss": 0.8692,
+      "step": 13180
+    },
+    {
+      "epoch": 0.8656275635767022,
+      "grad_norm": 0.3452875018119812,
+      "learning_rate": 0.0002,
+      "loss": 0.8429,
+      "step": 13190
+    },
+    {
+      "epoch": 0.8662838392124692,
+      "grad_norm": 0.36628788709640503,
+      "learning_rate": 0.0002,
+      "loss": 0.8872,
+      "step": 13200
+    },
+    {
+      "epoch": 0.8669401148482363,
+      "grad_norm": 0.4329487383365631,
+      "learning_rate": 0.0002,
+      "loss": 0.9141,
+      "step": 13210
+    },
+    {
+      "epoch": 0.8675963904840033,
+      "grad_norm": 0.36800137162208557,
+      "learning_rate": 0.0002,
+      "loss": 0.8586,
+      "step": 13220
+    },
+    {
+      "epoch": 0.8682526661197703,
+      "grad_norm": 0.5235224366188049,
+      "learning_rate": 0.0002,
+      "loss": 0.9133,
+      "step": 13230
+    },
+    {
+      "epoch": 0.8689089417555373,
+      "grad_norm": 0.40818873047828674,
+      "learning_rate": 0.0002,
+      "loss": 0.8826,
+      "step": 13240
+    },
+    {
+      "epoch": 0.8695652173913043,
+      "grad_norm": 0.46719685196876526,
+      "learning_rate": 0.0002,
+      "loss": 0.8869,
+      "step": 13250
+    },
+    {
+      "epoch": 0.8702214930270714,
+      "grad_norm": 0.34422767162323,
+      "learning_rate": 0.0002,
+      "loss": 0.9426,
+      "step": 13260
+    },
+    {
+      "epoch": 0.8708777686628384,
+      "grad_norm": 0.37454837560653687,
+      "learning_rate": 0.0002,
+      "loss": 0.8921,
+      "step": 13270
+    },
+    {
+      "epoch": 0.8715340442986054,
+      "grad_norm": 0.39750349521636963,
+      "learning_rate": 0.0002,
+      "loss": 0.89,
+      "step": 13280
+    },
+    {
+      "epoch": 0.8721903199343725,
+      "grad_norm": 0.3864808678627014,
+      "learning_rate": 0.0002,
+      "loss": 0.9256,
+      "step": 13290
+    },
+    {
+      "epoch": 0.8728465955701394,
+      "grad_norm": 0.36088764667510986,
+      "learning_rate": 0.0002,
+      "loss": 0.8663,
+      "step": 13300
+    },
+    {
+      "epoch": 0.8735028712059065,
+      "grad_norm": 0.384287029504776,
+      "learning_rate": 0.0002,
+      "loss": 0.8334,
+      "step": 13310
+    },
+    {
+      "epoch": 0.8741591468416735,
+      "grad_norm": 0.3988962173461914,
+      "learning_rate": 0.0002,
+      "loss": 0.8697,
+      "step": 13320
+    },
+    {
+      "epoch": 0.8748154224774405,
+      "grad_norm": 0.40126702189445496,
+      "learning_rate": 0.0002,
+      "loss": 0.8924,
+      "step": 13330
+    },
+    {
+      "epoch": 0.8754716981132076,
+      "grad_norm": 0.3931732475757599,
+      "learning_rate": 0.0002,
+      "loss": 0.8766,
+      "step": 13340
+    },
+    {
+      "epoch": 0.8761279737489746,
+      "grad_norm": 0.35348305106163025,
+      "learning_rate": 0.0002,
+      "loss": 0.9249,
+      "step": 13350
+    },
+    {
+      "epoch": 0.8767842493847416,
+      "grad_norm": 0.3603714108467102,
+      "learning_rate": 0.0002,
+      "loss": 0.9412,
+      "step": 13360
+    },
+    {
+      "epoch": 0.8774405250205086,
+      "grad_norm": 0.3853464126586914,
+      "learning_rate": 0.0002,
+      "loss": 0.8432,
+      "step": 13370
+    },
+    {
+      "epoch": 0.8780968006562756,
+      "grad_norm": 0.44406014680862427,
+      "learning_rate": 0.0002,
+      "loss": 0.9035,
+      "step": 13380
+    },
+    {
+      "epoch": 0.8787530762920427,
+      "grad_norm": 0.3563307225704193,
+      "learning_rate": 0.0002,
+      "loss": 0.8903,
+      "step": 13390
+    },
+    {
+      "epoch": 0.8794093519278097,
+      "grad_norm": 0.4507097005844116,
+      "learning_rate": 0.0002,
+      "loss": 0.9004,
+      "step": 13400
+    },
+    {
+      "epoch": 0.8800656275635766,
+      "grad_norm": 0.4069702923297882,
+      "learning_rate": 0.0002,
+      "loss": 0.9792,
+      "step": 13410
+    },
+    {
+      "epoch": 0.8807219031993437,
+      "grad_norm": 0.5091678500175476,
+      "learning_rate": 0.0002,
+      "loss": 0.8857,
+      "step": 13420
+    },
+    {
+      "epoch": 0.8813781788351107,
+      "grad_norm": 0.3784930408000946,
+      "learning_rate": 0.0002,
+      "loss": 0.9499,
+      "step": 13430
+    },
+    {
+      "epoch": 0.8820344544708778,
+      "grad_norm": 0.3937450051307678,
+      "learning_rate": 0.0002,
+      "loss": 0.9475,
+      "step": 13440
+    },
+    {
+      "epoch": 0.8826907301066448,
+      "grad_norm": 0.39602798223495483,
+      "learning_rate": 0.0002,
+      "loss": 0.9411,
+      "step": 13450
+    },
+    {
+      "epoch": 0.8833470057424118,
+      "grad_norm": 0.42562225461006165,
+      "learning_rate": 0.0002,
+      "loss": 0.947,
+      "step": 13460
+    },
+    {
+      "epoch": 0.8840032813781789,
+      "grad_norm": 0.3933939039707184,
+      "learning_rate": 0.0002,
+      "loss": 0.9748,
+      "step": 13470
+    },
+    {
+      "epoch": 0.8846595570139458,
+      "grad_norm": 0.4387489855289459,
+      "learning_rate": 0.0002,
+      "loss": 0.9206,
+      "step": 13480
+    },
+    {
+      "epoch": 0.8853158326497129,
+      "grad_norm": 0.3655209243297577,
+      "learning_rate": 0.0002,
+      "loss": 0.8417,
+      "step": 13490
+    },
+    {
+      "epoch": 0.8859721082854799,
+      "grad_norm": 0.40750762820243835,
+      "learning_rate": 0.0002,
+      "loss": 0.8601,
+      "step": 13500
+    },
+    {
+      "epoch": 0.8866283839212469,
+      "grad_norm": 0.5716604590415955,
+      "learning_rate": 0.0002,
+      "loss": 0.9343,
+      "step": 13510
+    },
+    {
+      "epoch": 0.887284659557014,
+      "grad_norm": 0.3286498785018921,
+      "learning_rate": 0.0002,
+      "loss": 0.8634,
+      "step": 13520
+    },
+    {
+      "epoch": 0.887940935192781,
+      "grad_norm": 0.4093165993690491,
+      "learning_rate": 0.0002,
+      "loss": 0.8875,
+      "step": 13530
+    },
+    {
+      "epoch": 0.888597210828548,
+      "grad_norm": 0.37128645181655884,
+      "learning_rate": 0.0002,
+      "loss": 0.9056,
+      "step": 13540
+    },
+    {
+      "epoch": 0.889253486464315,
+      "grad_norm": 0.9630060195922852,
+      "learning_rate": 0.0002,
+      "loss": 0.9356,
+      "step": 13550
+    },
+    {
+      "epoch": 0.889909762100082,
+      "grad_norm": 0.4119563698768616,
+      "learning_rate": 0.0002,
+      "loss": 0.9598,
+      "step": 13560
+    },
+    {
+      "epoch": 0.8905660377358491,
+      "grad_norm": 0.42105695605278015,
+      "learning_rate": 0.0002,
+      "loss": 0.8688,
+      "step": 13570
+    },
+    {
+      "epoch": 0.8912223133716161,
+      "grad_norm": 0.35517653822898865,
+      "learning_rate": 0.0002,
+      "loss": 0.9091,
+      "step": 13580
+    },
+    {
+      "epoch": 0.8918785890073831,
+      "grad_norm": 0.4380730986595154,
+      "learning_rate": 0.0002,
+      "loss": 0.9975,
+      "step": 13590
+    },
+    {
+      "epoch": 0.8925348646431501,
+      "grad_norm": 0.3875235915184021,
+      "learning_rate": 0.0002,
+      "loss": 0.8874,
+      "step": 13600
+    },
+    {
+      "epoch": 0.8931911402789171,
+      "grad_norm": 0.3194465935230255,
+      "learning_rate": 0.0002,
+      "loss": 0.8813,
+      "step": 13610
+    },
+    {
+      "epoch": 0.8938474159146842,
+      "grad_norm": 0.4448561370372772,
+      "learning_rate": 0.0002,
+      "loss": 0.8671,
+      "step": 13620
+    },
+    {
+      "epoch": 0.8945036915504512,
+      "grad_norm": 0.4257558286190033,
+      "learning_rate": 0.0002,
+      "loss": 0.88,
+      "step": 13630
+    },
+    {
+      "epoch": 0.8951599671862183,
+      "grad_norm": 0.3831070065498352,
+      "learning_rate": 0.0002,
+      "loss": 0.8844,
+      "step": 13640
+    },
+    {
+      "epoch": 0.8958162428219852,
+      "grad_norm": 0.3765697479248047,
+      "learning_rate": 0.0002,
+      "loss": 0.8836,
+      "step": 13650
+    },
+    {
+      "epoch": 0.8964725184577522,
+      "grad_norm": 0.3621887266635895,
+      "learning_rate": 0.0002,
+      "loss": 0.9015,
+      "step": 13660
+    },
+    {
+      "epoch": 0.8971287940935193,
+      "grad_norm": 0.4219911992549896,
+      "learning_rate": 0.0002,
+      "loss": 0.9147,
+      "step": 13670
+    },
+    {
+      "epoch": 0.8977850697292863,
+      "grad_norm": 0.550961971282959,
+      "learning_rate": 0.0002,
+      "loss": 0.878,
+      "step": 13680
+    },
+    {
+      "epoch": 0.8984413453650533,
+      "grad_norm": 0.41570132970809937,
+      "learning_rate": 0.0002,
+      "loss": 0.8807,
+      "step": 13690
+    },
+    {
+      "epoch": 0.8990976210008204,
+      "grad_norm": 0.34862181544303894,
+      "learning_rate": 0.0002,
+      "loss": 0.953,
+      "step": 13700
+    },
+    {
+      "epoch": 0.8997538966365873,
+      "grad_norm": 0.3568742871284485,
+      "learning_rate": 0.0002,
+      "loss": 0.8738,
+      "step": 13710
+    },
+    {
+      "epoch": 0.9004101722723544,
+      "grad_norm": 0.3789973258972168,
+      "learning_rate": 0.0002,
+      "loss": 0.9346,
+      "step": 13720
+    },
+    {
+      "epoch": 0.9010664479081214,
+      "grad_norm": 0.3775809109210968,
+      "learning_rate": 0.0002,
+      "loss": 0.8225,
+      "step": 13730
+    },
+    {
+      "epoch": 0.9017227235438884,
+      "grad_norm": 0.33509477972984314,
+      "learning_rate": 0.0002,
+      "loss": 0.9238,
+      "step": 13740
+    },
+    {
+      "epoch": 0.9023789991796555,
+      "grad_norm": 0.4410351514816284,
+      "learning_rate": 0.0002,
+      "loss": 0.915,
+      "step": 13750
+    },
+    {
+      "epoch": 0.9030352748154225,
+      "grad_norm": 0.44638893008232117,
+      "learning_rate": 0.0002,
+      "loss": 0.9066,
+      "step": 13760
+    },
+    {
+      "epoch": 0.9036915504511895,
+      "grad_norm": 0.38487187027931213,
+      "learning_rate": 0.0002,
+      "loss": 0.883,
+      "step": 13770
+    },
+    {
+      "epoch": 0.9043478260869565,
+      "grad_norm": 0.385796457529068,
+      "learning_rate": 0.0002,
+      "loss": 0.8886,
+      "step": 13780
+    },
+    {
+      "epoch": 0.9050041017227235,
+      "grad_norm": 0.4261656403541565,
+      "learning_rate": 0.0002,
+      "loss": 0.8781,
+      "step": 13790
+    },
+    {
+      "epoch": 0.9056603773584906,
+      "grad_norm": 0.38487741351127625,
+      "learning_rate": 0.0002,
+      "loss": 0.9237,
+      "step": 13800
+    },
+    {
+      "epoch": 0.9063166529942576,
+      "grad_norm": 0.40660005807876587,
+      "learning_rate": 0.0002,
+      "loss": 0.9357,
+      "step": 13810
+    },
+    {
+      "epoch": 0.9069729286300247,
+      "grad_norm": 0.4127330780029297,
+      "learning_rate": 0.0002,
+      "loss": 0.8943,
+      "step": 13820
+    },
+    {
+      "epoch": 0.9076292042657916,
+      "grad_norm": 0.4300757944583893,
+      "learning_rate": 0.0002,
+      "loss": 0.8981,
+      "step": 13830
+    },
+    {
+      "epoch": 0.9082854799015586,
+      "grad_norm": 0.3994467854499817,
+      "learning_rate": 0.0002,
+      "loss": 0.8956,
+      "step": 13840
+    },
+    {
+      "epoch": 0.9089417555373257,
+      "grad_norm": 0.4332261383533478,
+      "learning_rate": 0.0002,
+      "loss": 0.8949,
+      "step": 13850
+    },
+    {
+      "epoch": 0.9095980311730927,
+      "grad_norm": 0.3849696218967438,
+      "learning_rate": 0.0002,
+      "loss": 0.8897,
+      "step": 13860
+    },
+    {
+      "epoch": 0.9102543068088598,
+      "grad_norm": 0.39346274733543396,
+      "learning_rate": 0.0002,
+      "loss": 0.8897,
+      "step": 13870
+    },
+    {
+      "epoch": 0.9109105824446267,
+      "grad_norm": 0.39518049359321594,
+      "learning_rate": 0.0002,
+      "loss": 0.8847,
+      "step": 13880
+    },
+    {
+      "epoch": 0.9115668580803937,
+      "grad_norm": 0.4449180066585541,
+      "learning_rate": 0.0002,
+      "loss": 0.8982,
+      "step": 13890
+    },
+    {
+      "epoch": 0.9122231337161608,
+      "grad_norm": 0.41052138805389404,
+      "learning_rate": 0.0002,
+      "loss": 0.9057,
+      "step": 13900
+    },
+    {
+      "epoch": 0.9128794093519278,
+      "grad_norm": 0.36827564239501953,
+      "learning_rate": 0.0002,
+      "loss": 0.905,
+      "step": 13910
+    },
+    {
+      "epoch": 0.9135356849876949,
+      "grad_norm": 0.3875851631164551,
+      "learning_rate": 0.0002,
+      "loss": 0.9484,
+      "step": 13920
+    },
+    {
+      "epoch": 0.9141919606234619,
+      "grad_norm": 0.402854859828949,
+      "learning_rate": 0.0002,
+      "loss": 0.9125,
+      "step": 13930
+    },
+    {
+      "epoch": 0.9148482362592288,
+      "grad_norm": 0.3584592938423157,
+      "learning_rate": 0.0002,
+      "loss": 0.9014,
+      "step": 13940
+    },
+    {
+      "epoch": 0.9155045118949959,
+      "grad_norm": 0.3516979515552521,
+      "learning_rate": 0.0002,
+      "loss": 0.8432,
+      "step": 13950
+    },
+    {
+      "epoch": 0.9161607875307629,
+      "grad_norm": 0.4411509335041046,
+      "learning_rate": 0.0002,
+      "loss": 0.8747,
+      "step": 13960
+    },
+    {
+      "epoch": 0.9168170631665299,
+      "grad_norm": 0.47830596566200256,
+      "learning_rate": 0.0002,
+      "loss": 0.8758,
+      "step": 13970
+    },
+    {
+      "epoch": 0.917473338802297,
+      "grad_norm": 0.3669400215148926,
+      "learning_rate": 0.0002,
+      "loss": 0.915,
+      "step": 13980
+    },
+    {
+      "epoch": 0.918129614438064,
+      "grad_norm": 0.4361213147640228,
+      "learning_rate": 0.0002,
+      "loss": 0.9265,
+      "step": 13990
+    },
+    {
+      "epoch": 0.918785890073831,
+      "grad_norm": 0.4224131405353546,
+      "learning_rate": 0.0002,
+      "loss": 0.9106,
+      "step": 14000
+    },
+    {
+      "epoch": 0.919442165709598,
+      "grad_norm": 0.5785587430000305,
+      "learning_rate": 0.0002,
+      "loss": 0.9138,
+      "step": 14010
+    },
+    {
+      "epoch": 0.920098441345365,
+      "grad_norm": 0.40788379311561584,
+      "learning_rate": 0.0002,
+      "loss": 0.9078,
+      "step": 14020
+    },
+    {
+      "epoch": 0.9207547169811321,
+      "grad_norm": 0.40879732370376587,
+      "learning_rate": 0.0002,
+      "loss": 0.9241,
+      "step": 14030
+    },
+    {
+      "epoch": 0.9214109926168991,
+      "grad_norm": 0.4031982421875,
+      "learning_rate": 0.0002,
+      "loss": 0.8565,
+      "step": 14040
+    },
+    {
+      "epoch": 0.9220672682526662,
+      "grad_norm": 0.3457014560699463,
+      "learning_rate": 0.0002,
+      "loss": 0.8224,
+      "step": 14050
+    },
+    {
+      "epoch": 0.9227235438884331,
+      "grad_norm": 0.38608697056770325,
+      "learning_rate": 0.0002,
+      "loss": 0.89,
+      "step": 14060
+    },
+    {
+      "epoch": 0.9233798195242001,
+      "grad_norm": 0.39772507548332214,
+      "learning_rate": 0.0002,
+      "loss": 0.8637,
+      "step": 14070
+    },
+    {
+      "epoch": 0.9240360951599672,
+      "grad_norm": 0.3684687316417694,
+      "learning_rate": 0.0002,
+      "loss": 0.9102,
+      "step": 14080
+    },
+    {
+      "epoch": 0.9246923707957342,
+      "grad_norm": 0.4479428827762604,
+      "learning_rate": 0.0002,
+      "loss": 0.886,
+      "step": 14090
+    },
+    {
+      "epoch": 0.9253486464315013,
+      "grad_norm": 0.39918506145477295,
+      "learning_rate": 0.0002,
+      "loss": 0.857,
+      "step": 14100
+    },
+    {
+      "epoch": 0.9260049220672683,
+      "grad_norm": 0.4163782000541687,
+      "learning_rate": 0.0002,
+      "loss": 0.8765,
+      "step": 14110
+    },
+    {
+      "epoch": 0.9266611977030352,
+      "grad_norm": 0.40232136845588684,
+      "learning_rate": 0.0002,
+      "loss": 0.9022,
+      "step": 14120
+    },
+    {
+      "epoch": 0.9273174733388023,
+      "grad_norm": 0.36188459396362305,
+      "learning_rate": 0.0002,
+      "loss": 0.8815,
+      "step": 14130
+    },
+    {
+      "epoch": 0.9279737489745693,
+      "grad_norm": 0.42056623101234436,
+      "learning_rate": 0.0002,
+      "loss": 0.9072,
+      "step": 14140
+    },
+    {
+      "epoch": 0.9286300246103364,
+      "grad_norm": 0.3129708468914032,
+      "learning_rate": 0.0002,
+      "loss": 0.8678,
+      "step": 14150
+    },
+    {
+      "epoch": 0.9292863002461034,
+      "grad_norm": 0.4068623185157776,
+      "learning_rate": 0.0002,
+      "loss": 0.8895,
+      "step": 14160
+    },
+    {
+      "epoch": 0.9299425758818703,
+      "grad_norm": 0.43788865208625793,
+      "learning_rate": 0.0002,
+      "loss": 0.8939,
+      "step": 14170
+    },
+    {
+      "epoch": 0.9305988515176374,
+      "grad_norm": 0.35850921273231506,
+      "learning_rate": 0.0002,
+      "loss": 0.8954,
+      "step": 14180
+    },
+    {
+      "epoch": 0.9312551271534044,
+      "grad_norm": 0.41121411323547363,
+      "learning_rate": 0.0002,
+      "loss": 0.9001,
+      "step": 14190
+    },
+    {
+      "epoch": 0.9319114027891715,
+      "grad_norm": 0.4249218702316284,
+      "learning_rate": 0.0002,
+      "loss": 0.866,
+      "step": 14200
+    },
+    {
+      "epoch": 0.9325676784249385,
+      "grad_norm": 0.3779831826686859,
+      "learning_rate": 0.0002,
+      "loss": 0.9195,
+      "step": 14210
+    },
+    {
+      "epoch": 0.9332239540607055,
+      "grad_norm": 0.36781951785087585,
+      "learning_rate": 0.0002,
+      "loss": 0.8436,
+      "step": 14220
+    },
+    {
+      "epoch": 0.9338802296964726,
+      "grad_norm": 0.36741000413894653,
+      "learning_rate": 0.0002,
+      "loss": 0.8976,
+      "step": 14230
+    },
+    {
+      "epoch": 0.9345365053322395,
+      "grad_norm": 0.43414175510406494,
+      "learning_rate": 0.0002,
+      "loss": 0.9137,
+      "step": 14240
+    },
+    {
+      "epoch": 0.9351927809680065,
+      "grad_norm": 0.4406278431415558,
+      "learning_rate": 0.0002,
+      "loss": 0.8648,
+      "step": 14250
+    },
+    {
+      "epoch": 0.9358490566037736,
+      "grad_norm": 0.386152982711792,
+      "learning_rate": 0.0002,
+      "loss": 0.8705,
+      "step": 14260
+    },
+    {
+      "epoch": 0.9365053322395406,
+      "grad_norm": 0.3971618711948395,
+      "learning_rate": 0.0002,
+      "loss": 0.8697,
+      "step": 14270
+    },
+    {
+      "epoch": 0.9371616078753077,
+      "grad_norm": 0.3366684317588806,
+      "learning_rate": 0.0002,
+      "loss": 0.8925,
+      "step": 14280
+    },
+    {
+      "epoch": 0.9378178835110746,
+      "grad_norm": 0.42566195130348206,
+      "learning_rate": 0.0002,
+      "loss": 0.8702,
+      "step": 14290
+    },
+    {
+      "epoch": 0.9384741591468416,
+      "grad_norm": 0.3734602928161621,
+      "learning_rate": 0.0002,
+      "loss": 0.8767,
+      "step": 14300
+    },
+    {
+      "epoch": 0.9391304347826087,
+      "grad_norm": 0.4279540479183197,
+      "learning_rate": 0.0002,
+      "loss": 0.8889,
+      "step": 14310
+    },
+    {
+      "epoch": 0.9397867104183757,
+      "grad_norm": 0.43050041794776917,
+      "learning_rate": 0.0002,
+      "loss": 0.9319,
+      "step": 14320
+    },
+    {
+      "epoch": 0.9404429860541428,
+      "grad_norm": 0.41909968852996826,
+      "learning_rate": 0.0002,
+      "loss": 0.91,
+      "step": 14330
+    },
+    {
+      "epoch": 0.9410992616899098,
+      "grad_norm": 0.39651772379875183,
+      "learning_rate": 0.0002,
+      "loss": 0.8782,
+      "step": 14340
+    },
+    {
+      "epoch": 0.9417555373256767,
+      "grad_norm": 0.4171423614025116,
+      "learning_rate": 0.0002,
+      "loss": 0.8868,
+      "step": 14350
+    },
+    {
+      "epoch": 0.9424118129614438,
+      "grad_norm": 0.44906023144721985,
+      "learning_rate": 0.0002,
+      "loss": 0.8751,
+      "step": 14360
+    },
+    {
+      "epoch": 0.9430680885972108,
+      "grad_norm": 0.4213627576828003,
+      "learning_rate": 0.0002,
+      "loss": 0.9048,
+      "step": 14370
+    },
+    {
+      "epoch": 0.9437243642329779,
+      "grad_norm": 0.38457417488098145,
+      "learning_rate": 0.0002,
+      "loss": 0.9257,
+      "step": 14380
+    },
+    {
+      "epoch": 0.9443806398687449,
+      "grad_norm": 0.43104225397109985,
+      "learning_rate": 0.0002,
+      "loss": 0.8421,
+      "step": 14390
+    },
+    {
+      "epoch": 0.9450369155045119,
+      "grad_norm": 0.40090736746788025,
+      "learning_rate": 0.0002,
+      "loss": 0.865,
+      "step": 14400
+    },
+    {
+      "epoch": 0.9456931911402789,
+      "grad_norm": 0.36180031299591064,
+      "learning_rate": 0.0002,
+      "loss": 0.8746,
+      "step": 14410
+    },
+    {
+      "epoch": 0.9463494667760459,
+      "grad_norm": 0.4608926475048065,
+      "learning_rate": 0.0002,
+      "loss": 0.9297,
+      "step": 14420
+    },
+    {
+      "epoch": 0.947005742411813,
+      "grad_norm": 0.44056418538093567,
+      "learning_rate": 0.0002,
+      "loss": 0.9131,
+      "step": 14430
+    },
+    {
+      "epoch": 0.94766201804758,
+      "grad_norm": 0.334051251411438,
+      "learning_rate": 0.0002,
+      "loss": 0.8368,
+      "step": 14440
+    },
+    {
+      "epoch": 0.948318293683347,
+      "grad_norm": 0.45580169558525085,
+      "learning_rate": 0.0002,
+      "loss": 0.8557,
+      "step": 14450
+    },
+    {
+      "epoch": 0.9489745693191141,
+      "grad_norm": 0.3898446261882782,
+      "learning_rate": 0.0002,
+      "loss": 0.8828,
+      "step": 14460
+    },
+    {
+      "epoch": 0.949630844954881,
+      "grad_norm": 0.465259850025177,
+      "learning_rate": 0.0002,
+      "loss": 0.9436,
+      "step": 14470
+    },
+    {
+      "epoch": 0.9502871205906481,
+      "grad_norm": 0.3595256805419922,
+      "learning_rate": 0.0002,
+      "loss": 0.8907,
+      "step": 14480
+    },
+    {
+      "epoch": 0.9509433962264151,
+      "grad_norm": 0.3710390031337738,
+      "learning_rate": 0.0002,
+      "loss": 0.8593,
+      "step": 14490
+    },
+    {
+      "epoch": 0.9515996718621821,
+      "grad_norm": 0.4002886712551117,
+      "learning_rate": 0.0002,
+      "loss": 0.8373,
+      "step": 14500
+    },
+    {
+      "epoch": 0.9522559474979492,
+      "grad_norm": 0.4221348464488983,
+      "learning_rate": 0.0002,
+      "loss": 0.9091,
+      "step": 14510
+    },
+    {
+      "epoch": 0.9529122231337162,
+      "grad_norm": 0.4163874089717865,
+      "learning_rate": 0.0002,
+      "loss": 0.9205,
+      "step": 14520
+    },
+    {
+      "epoch": 0.9535684987694831,
+      "grad_norm": 0.40220746397972107,
+      "learning_rate": 0.0002,
+      "loss": 0.9125,
+      "step": 14530
+    },
+    {
+      "epoch": 0.9542247744052502,
+      "grad_norm": 0.45836037397384644,
+      "learning_rate": 0.0002,
+      "loss": 0.854,
+      "step": 14540
+    },
+    {
+      "epoch": 0.9548810500410172,
+      "grad_norm": 0.420478492975235,
+      "learning_rate": 0.0002,
+      "loss": 0.911,
+      "step": 14550
+    },
+    {
+      "epoch": 0.9555373256767843,
+      "grad_norm": 0.45627933740615845,
+      "learning_rate": 0.0002,
+      "loss": 0.8932,
+      "step": 14560
+    },
+    {
+      "epoch": 0.9561936013125513,
+      "grad_norm": 0.41290518641471863,
+      "learning_rate": 0.0002,
+      "loss": 0.8926,
+      "step": 14570
+    },
+    {
+      "epoch": 0.9568498769483182,
+      "grad_norm": 0.46660760045051575,
+      "learning_rate": 0.0002,
+      "loss": 0.8936,
+      "step": 14580
+    },
+    {
+      "epoch": 0.9575061525840853,
+      "grad_norm": 0.45327240228652954,
+      "learning_rate": 0.0002,
+      "loss": 0.9132,
+      "step": 14590
+    },
+    {
+      "epoch": 0.9581624282198523,
+      "grad_norm": 0.41973528265953064,
+      "learning_rate": 0.0002,
+      "loss": 0.8601,
+      "step": 14600
+    },
+    {
+      "epoch": 0.9588187038556194,
+      "grad_norm": 0.4022239148616791,
+      "learning_rate": 0.0002,
+      "loss": 0.8278,
+      "step": 14610
+    },
+    {
+      "epoch": 0.9594749794913864,
+      "grad_norm": 0.3226695954799652,
+      "learning_rate": 0.0002,
+      "loss": 0.8514,
+      "step": 14620
+    },
+    {
+      "epoch": 0.9601312551271534,
+      "grad_norm": 0.4233718812465668,
+      "learning_rate": 0.0002,
+      "loss": 0.8526,
+      "step": 14630
+    },
+    {
+      "epoch": 0.9607875307629204,
+      "grad_norm": 0.37561315298080444,
+      "learning_rate": 0.0002,
+      "loss": 0.9009,
+      "step": 14640
+    },
+    {
+      "epoch": 0.9614438063986874,
+      "grad_norm": 0.44843146204948425,
+      "learning_rate": 0.0002,
+      "loss": 0.8854,
+      "step": 14650
+    },
+    {
+      "epoch": 0.9621000820344545,
+      "grad_norm": 0.40062573552131653,
+      "learning_rate": 0.0002,
+      "loss": 0.8863,
+      "step": 14660
+    },
+    {
+      "epoch": 0.9627563576702215,
+      "grad_norm": 0.4720284938812256,
+      "learning_rate": 0.0002,
+      "loss": 0.8912,
+      "step": 14670
+    },
+    {
+      "epoch": 0.9634126333059885,
+      "grad_norm": 0.43565067648887634,
+      "learning_rate": 0.0002,
+      "loss": 0.89,
+      "step": 14680
+    },
+    {
+      "epoch": 0.9640689089417556,
+      "grad_norm": 0.41181448101997375,
+      "learning_rate": 0.0002,
+      "loss": 0.8612,
+      "step": 14690
+    },
+    {
+      "epoch": 0.9647251845775225,
+      "grad_norm": 0.4344978630542755,
+      "learning_rate": 0.0002,
+      "loss": 0.9464,
+      "step": 14700
+    },
+    {
+      "epoch": 0.9653814602132896,
+      "grad_norm": 0.42558008432388306,
+      "learning_rate": 0.0002,
+      "loss": 0.855,
+      "step": 14710
+    },
+    {
+      "epoch": 0.9660377358490566,
+      "grad_norm": 0.39161136746406555,
+      "learning_rate": 0.0002,
+      "loss": 0.8945,
+      "step": 14720
+    },
+    {
+      "epoch": 0.9666940114848236,
+      "grad_norm": 0.3784191608428955,
+      "learning_rate": 0.0002,
+      "loss": 0.8587,
+      "step": 14730
+    },
+    {
+      "epoch": 0.9673502871205907,
+      "grad_norm": 0.39039477705955505,
+      "learning_rate": 0.0002,
+      "loss": 0.8566,
+      "step": 14740
+    },
+    {
+      "epoch": 0.9680065627563577,
+      "grad_norm": 0.5566018223762512,
+      "learning_rate": 0.0002,
+      "loss": 0.9006,
+      "step": 14750
+    },
+    {
+      "epoch": 0.9686628383921247,
+      "grad_norm": 0.38877877593040466,
+      "learning_rate": 0.0002,
+      "loss": 0.9222,
+      "step": 14760
+    },
+    {
+      "epoch": 0.9693191140278917,
+      "grad_norm": 0.33369940519332886,
+      "learning_rate": 0.0002,
+      "loss": 0.8811,
+      "step": 14770
+    },
+    {
+      "epoch": 0.9699753896636587,
+      "grad_norm": 0.5109888315200806,
+      "learning_rate": 0.0002,
+      "loss": 0.8743,
+      "step": 14780
+    },
+    {
+      "epoch": 0.9706316652994258,
+      "grad_norm": 0.5364375114440918,
+      "learning_rate": 0.0002,
+      "loss": 0.8883,
+      "step": 14790
+    },
+    {
+      "epoch": 0.9712879409351928,
+      "grad_norm": 0.3702435791492462,
+      "learning_rate": 0.0002,
+      "loss": 0.8696,
+      "step": 14800
+    },
+    {
+      "epoch": 0.9719442165709598,
+      "grad_norm": 0.5094677805900574,
+      "learning_rate": 0.0002,
+      "loss": 0.9144,
+      "step": 14810
+    },
+    {
+      "epoch": 0.9726004922067268,
+      "grad_norm": 0.3565915524959564,
+      "learning_rate": 0.0002,
+      "loss": 0.8823,
+      "step": 14820
+    },
+    {
+      "epoch": 0.9732567678424938,
+      "grad_norm": 0.44756242632865906,
+      "learning_rate": 0.0002,
+      "loss": 0.8964,
+      "step": 14830
+    },
+    {
+      "epoch": 0.9739130434782609,
+      "grad_norm": 0.4272070527076721,
+      "learning_rate": 0.0002,
+      "loss": 0.8524,
+      "step": 14840
+    },
+    {
+      "epoch": 0.9745693191140279,
+      "grad_norm": 0.4303967356681824,
+      "learning_rate": 0.0002,
+      "loss": 0.8551,
+      "step": 14850
+    },
+    {
+      "epoch": 0.9752255947497949,
+      "grad_norm": 0.4528018534183502,
+      "learning_rate": 0.0002,
+      "loss": 0.9303,
+      "step": 14860
+    },
+    {
+      "epoch": 0.975881870385562,
+      "grad_norm": 0.5745970606803894,
+      "learning_rate": 0.0002,
+      "loss": 0.8465,
+      "step": 14870
+    },
+    {
+      "epoch": 0.9765381460213289,
+      "grad_norm": 0.43847736716270447,
+      "learning_rate": 0.0002,
+      "loss": 0.9496,
+      "step": 14880
+    },
+    {
+      "epoch": 0.977194421657096,
+      "grad_norm": 0.4512104094028473,
+      "learning_rate": 0.0002,
+      "loss": 0.8647,
+      "step": 14890
+    },
+    {
+      "epoch": 0.977850697292863,
+      "grad_norm": 0.4573594331741333,
+      "learning_rate": 0.0002,
+      "loss": 0.9712,
+      "step": 14900
+    },
+    {
+      "epoch": 0.97850697292863,
+      "grad_norm": 0.4297037720680237,
+      "learning_rate": 0.0002,
+      "loss": 0.898,
+      "step": 14910
+    },
+    {
+      "epoch": 0.9791632485643971,
+      "grad_norm": 0.3769957721233368,
+      "learning_rate": 0.0002,
+      "loss": 0.9115,
+      "step": 14920
+    },
+    {
+      "epoch": 0.979819524200164,
+      "grad_norm": 0.3524457514286041,
+      "learning_rate": 0.0002,
+      "loss": 0.8956,
+      "step": 14930
+    },
+    {
+      "epoch": 0.9804757998359311,
+      "grad_norm": 0.3965851664543152,
+      "learning_rate": 0.0002,
+      "loss": 0.9254,
+      "step": 14940
+    },
+    {
+      "epoch": 0.9811320754716981,
+      "grad_norm": 0.3593656122684479,
+      "learning_rate": 0.0002,
+      "loss": 0.8869,
+      "step": 14950
+    },
+    {
+      "epoch": 0.9817883511074651,
+      "grad_norm": 0.35391807556152344,
+      "learning_rate": 0.0002,
+      "loss": 0.8859,
+      "step": 14960
+    },
+    {
+      "epoch": 0.9824446267432322,
+      "grad_norm": 0.40651339292526245,
+      "learning_rate": 0.0002,
+      "loss": 0.9111,
+      "step": 14970
+    },
+    {
+      "epoch": 0.9831009023789992,
+      "grad_norm": 0.4370724558830261,
+      "learning_rate": 0.0002,
+      "loss": 0.8714,
+      "step": 14980
+    },
+    {
+      "epoch": 0.9837571780147663,
+      "grad_norm": 0.37859413027763367,
+      "learning_rate": 0.0002,
+      "loss": 0.8776,
+      "step": 14990
+    },
+    {
+      "epoch": 0.9844134536505332,
+      "grad_norm": 0.4329196512699127,
+      "learning_rate": 0.0002,
+      "loss": 0.8744,
+      "step": 15000
+    },
+    {
+      "epoch": 0.9850697292863002,
+      "grad_norm": 0.4119299054145813,
+      "learning_rate": 0.0002,
+      "loss": 0.8384,
+      "step": 15010
+    },
+    {
+      "epoch": 0.9857260049220673,
+      "grad_norm": 0.4084014892578125,
+      "learning_rate": 0.0002,
+      "loss": 0.8885,
+      "step": 15020
+    },
+    {
+      "epoch": 0.9863822805578343,
+      "grad_norm": 0.34770357608795166,
+      "learning_rate": 0.0002,
+      "loss": 0.9255,
+      "step": 15030
+    },
+    {
+      "epoch": 0.9870385561936013,
+      "grad_norm": 0.429995596408844,
+      "learning_rate": 0.0002,
+      "loss": 0.9074,
+      "step": 15040
+    },
+    {
+      "epoch": 0.9876948318293683,
+      "grad_norm": 0.4588816165924072,
+      "learning_rate": 0.0002,
+      "loss": 0.8251,
+      "step": 15050
+    },
+    {
+      "epoch": 0.9883511074651353,
+      "grad_norm": 0.47414910793304443,
+      "learning_rate": 0.0002,
+      "loss": 0.895,
+      "step": 15060
+    },
+    {
+      "epoch": 0.9890073831009024,
+      "grad_norm": 0.3370365798473358,
+      "learning_rate": 0.0002,
+      "loss": 0.8688,
+      "step": 15070
+    },
+    {
+      "epoch": 0.9896636587366694,
+      "grad_norm": 0.3697716295719147,
+      "learning_rate": 0.0002,
+      "loss": 0.8775,
+      "step": 15080
+    },
+    {
+      "epoch": 0.9903199343724364,
+      "grad_norm": 0.31965479254722595,
+      "learning_rate": 0.0002,
+      "loss": 0.8949,
+      "step": 15090
+    },
+    {
+      "epoch": 0.9909762100082035,
+      "grad_norm": 0.5081075429916382,
+      "learning_rate": 0.0002,
+      "loss": 0.9519,
+      "step": 15100
+    },
+    {
+      "epoch": 0.9916324856439704,
+      "grad_norm": 0.44397613406181335,
+      "learning_rate": 0.0002,
+      "loss": 0.9012,
+      "step": 15110
+    },
+    {
+      "epoch": 0.9922887612797375,
+      "grad_norm": 0.30696988105773926,
+      "learning_rate": 0.0002,
+      "loss": 0.8327,
+      "step": 15120
+    },
+    {
+      "epoch": 0.9929450369155045,
+      "grad_norm": 0.4071432650089264,
+      "learning_rate": 0.0002,
+      "loss": 0.8785,
+      "step": 15130
+    },
+    {
+      "epoch": 0.9936013125512715,
+      "grad_norm": 0.5356084108352661,
+      "learning_rate": 0.0002,
+      "loss": 0.8879,
+      "step": 15140
+    },
+    {
+      "epoch": 0.9942575881870386,
+      "grad_norm": 0.36654597520828247,
+      "learning_rate": 0.0002,
+      "loss": 0.8593,
+      "step": 15150
+    },
+    {
+      "epoch": 0.9949138638228056,
+      "grad_norm": 0.38214483857154846,
+      "learning_rate": 0.0002,
+      "loss": 0.92,
+      "step": 15160
+    },
+    {
+      "epoch": 0.9955701394585726,
+      "grad_norm": 0.4340892434120178,
+      "learning_rate": 0.0002,
+      "loss": 0.9101,
+      "step": 15170
+    },
+    {
+      "epoch": 0.9962264150943396,
+      "grad_norm": 0.41310828924179077,
+      "learning_rate": 0.0002,
+      "loss": 0.9049,
+      "step": 15180
+    },
+    {
+      "epoch": 0.9968826907301066,
+      "grad_norm": 0.4932044744491577,
+      "learning_rate": 0.0002,
+      "loss": 0.8557,
+      "step": 15190
+    },
+    {
+      "epoch": 0.9975389663658737,
+      "grad_norm": 0.45371273159980774,
+      "learning_rate": 0.0002,
+      "loss": 0.8989,
+      "step": 15200
+    },
+    {
+      "epoch": 0.9981952420016407,
+      "grad_norm": 0.42956778407096863,
+      "learning_rate": 0.0002,
+      "loss": 0.9003,
+      "step": 15210
+    },
+    {
+      "epoch": 0.9988515176374078,
+      "grad_norm": 0.4343477487564087,
+      "learning_rate": 0.0002,
+      "loss": 0.8763,
+      "step": 15220
+    },
+    {
+      "epoch": 0.9995077932731747,
+      "grad_norm": 0.4425382912158966,
+      "learning_rate": 0.0002,
+      "loss": 0.8832,
+      "step": 15230
+    },
+    {
+      "epoch": 0.9999671862182117,
+      "eval_loss": 1.055190086364746,
+      "eval_runtime": 110.4581,
+      "eval_samples_per_second": 6.564,
+      "eval_steps_per_second": 0.824,
+      "step": 15237
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 121896,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 8,
+  "save_steps": 200,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 7.82655074009088e+17,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}
diff --git a/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/training_args.bin b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/training_args.bin
new file mode 100644
index 0000000000000000000000000000000000000000..33b12debaac07d7469e3dd63cd83a73ab48adf47
--- /dev/null
+++ b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237/training_args.bin
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fe28e20a8c72ab0a47191422d1378f55aaa44a84900b7a9377d408d348d306e6
+size 5560
diff --git a/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/training_log.jsonl b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/training_log.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..206ba0d460b961bd136272c62fa1db52149b9411
--- /dev/null
+++ b/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/training_log.jsonl
@@ -0,0 +1,2 @@
+{"epoch": 0.9999671862182117, "step": 15237, "epoch_duration": 41000.40024757385, "total_accumulated_duration": 41000.40024757385, "gpu_info": {"GPU_0": "NVIDIA A100-PCIE-40GB"}, "memory_usage": {"avg_memory_usage": {"GPU_0": 7887.97119140625}, "peak_memory_usage": {"GPU_0": 11696.9921875}, "avg_memory_reserved": {"GPU_0": 12786.0}, "peak_memory_reserved": {"GPU_0": 12786.0}, "total_memory": {"GPU_0": 40444.375}}, "best_checkpoint_path": "N/A", "params": {"epochs": 8, "batch_size": 1, "learning_rate": 0.0002, "gradient_accumulation_steps": 2, "warmup_ratio": 0.03, "max_grad_norm": 0.3, "lora_alpha": 32, "lora_dropout": 0.05, "lora_r": 64, "tf32": true, "seed": 42}, "log_history": [{"loss": 2.6059, "grad_norm": 1.0711857080459595, "learning_rate": 0.0002, "epoch": 0.0006562756357670221, "step": 10}, {"loss": 1.7643, "grad_norm": 0.935492992401123, "learning_rate": 0.0002, "epoch": 0.0013125512715340443, "step": 20}, {"loss": 1.2573, "grad_norm": 0.908809244632721, "learning_rate": 0.0002, "epoch": 0.0019688269073010667, "step": 30}, {"loss": 1.1011, "grad_norm": 0.3497907221317291, "learning_rate": 0.0002, "epoch": 0.0026251025430680886, "step": 40}, {"loss": 1.0866, "grad_norm": 0.32830339670181274, "learning_rate": 0.0002, "epoch": 0.003281378178835111, "step": 50}, {"loss": 1.0104, "grad_norm": 0.2850394546985626, "learning_rate": 0.0002, "epoch": 0.003937653814602133, "step": 60}, {"loss": 1.082, "grad_norm": 0.3804827928543091, "learning_rate": 0.0002, "epoch": 0.004593929450369155, "step": 70}, {"loss": 1.052, "grad_norm": 0.23506930470466614, "learning_rate": 0.0002, "epoch": 0.005250205086136177, "step": 80}, {"loss": 1.0593, "grad_norm": 0.2644859552383423, "learning_rate": 0.0002, "epoch": 0.005906480721903199, "step": 90}, {"loss": 0.9978, "grad_norm": 0.36523646116256714, "learning_rate": 0.0002, "epoch": 0.006562756357670222, "step": 100}, {"loss": 0.9958, "grad_norm": 0.3195570111274719, "learning_rate": 0.0002, "epoch": 0.007219031993437244, "step": 110}, {"loss": 1.0041, "grad_norm": 0.2886694371700287, "learning_rate": 0.0002, "epoch": 0.007875307629204267, "step": 120}, {"loss": 1.0623, "grad_norm": 0.3913154900074005, "learning_rate": 0.0002, "epoch": 0.008531583264971289, "step": 130}, {"loss": 1.1128, "grad_norm": 0.3181937336921692, "learning_rate": 0.0002, "epoch": 0.00918785890073831, "step": 140}, {"loss": 0.9989, "grad_norm": 0.2629619538784027, "learning_rate": 0.0002, "epoch": 0.009844134536505332, "step": 150}, {"loss": 1.0226, "grad_norm": 0.30438563227653503, "learning_rate": 0.0002, "epoch": 0.010500410172272354, "step": 160}, {"loss": 1.0321, "grad_norm": 0.2876931130886078, "learning_rate": 0.0002, "epoch": 0.011156685808039376, "step": 170}, {"loss": 0.9824, "grad_norm": 0.29188141226768494, "learning_rate": 0.0002, "epoch": 0.011812961443806398, "step": 180}, {"loss": 0.9439, "grad_norm": 0.2645126283168793, "learning_rate": 0.0002, "epoch": 0.01246923707957342, "step": 190}, {"loss": 1.0273, "grad_norm": 0.26031428575515747, "learning_rate": 0.0002, "epoch": 0.013125512715340444, "step": 200}, {"loss": 1.0518, "grad_norm": 0.25812748074531555, "learning_rate": 0.0002, "epoch": 0.013781788351107466, "step": 210}, {"loss": 0.9502, "grad_norm": 0.24913132190704346, "learning_rate": 0.0002, "epoch": 0.014438063986874488, "step": 220}, {"loss": 1.0131, "grad_norm": 0.30332663655281067, "learning_rate": 0.0002, "epoch": 0.01509433962264151, "step": 230}, {"loss": 1.0459, "grad_norm": 0.25207284092903137, "learning_rate": 0.0002, "epoch": 0.015750615258408533, "step": 240}, {"loss": 0.9798, "grad_norm": 0.26384010910987854, "learning_rate": 0.0002, "epoch": 0.016406890894175553, "step": 250}, {"loss": 1.061, "grad_norm": 0.28651612997055054, "learning_rate": 0.0002, "epoch": 0.017063166529942577, "step": 260}, {"loss": 0.991, "grad_norm": 0.2879799008369446, "learning_rate": 0.0002, "epoch": 0.017719442165709597, "step": 270}, {"loss": 0.9643, "grad_norm": 0.28661131858825684, "learning_rate": 0.0002, "epoch": 0.01837571780147662, "step": 280}, {"loss": 0.9903, "grad_norm": 0.265348345041275, "learning_rate": 0.0002, "epoch": 0.01903199343724364, "step": 290}, {"loss": 1.011, "grad_norm": 0.24186863005161285, "learning_rate": 0.0002, "epoch": 0.019688269073010665, "step": 300}, {"loss": 1.0245, "grad_norm": 0.25996068120002747, "learning_rate": 0.0002, "epoch": 0.020344544708777685, "step": 310}, {"loss": 0.9987, "grad_norm": 0.29615098237991333, "learning_rate": 0.0002, "epoch": 0.02100082034454471, "step": 320}, {"loss": 1.04, "grad_norm": 0.2429388016462326, "learning_rate": 0.0002, "epoch": 0.021657095980311732, "step": 330}, {"loss": 0.9703, "grad_norm": 0.26882505416870117, "learning_rate": 0.0002, "epoch": 0.022313371616078752, "step": 340}, {"loss": 0.9686, "grad_norm": 0.283328652381897, "learning_rate": 0.0002, "epoch": 0.022969647251845776, "step": 350}, {"loss": 0.952, "grad_norm": 0.3115910589694977, "learning_rate": 0.0002, "epoch": 0.023625922887612796, "step": 360}, {"loss": 1.0112, "grad_norm": 0.27969497442245483, "learning_rate": 0.0002, "epoch": 0.02428219852337982, "step": 370}, {"loss": 1.0618, "grad_norm": 0.30471885204315186, "learning_rate": 0.0002, "epoch": 0.02493847415914684, "step": 380}, {"loss": 1.0189, "grad_norm": 0.3183926045894623, "learning_rate": 0.0002, "epoch": 0.025594749794913864, "step": 390}, {"loss": 1.0148, "grad_norm": 0.27311646938323975, "learning_rate": 0.0002, "epoch": 0.026251025430680888, "step": 400}, {"loss": 0.9672, "grad_norm": 0.29148945212364197, "learning_rate": 0.0002, "epoch": 0.026907301066447908, "step": 410}, {"loss": 0.9473, "grad_norm": 0.2386617809534073, "learning_rate": 0.0002, "epoch": 0.02756357670221493, "step": 420}, {"loss": 1.0722, "grad_norm": 0.2546529471874237, "learning_rate": 0.0002, "epoch": 0.02821985233798195, "step": 430}, {"loss": 1.0017, "grad_norm": 0.27932611107826233, "learning_rate": 0.0002, "epoch": 0.028876127973748975, "step": 440}, {"loss": 0.9988, "grad_norm": 0.31259334087371826, "learning_rate": 0.0002, "epoch": 0.029532403609515995, "step": 450}, {"loss": 1.0404, "grad_norm": 0.2675893008708954, "learning_rate": 0.0002, "epoch": 0.03018867924528302, "step": 460}, {"loss": 1.0018, "grad_norm": 0.24047039449214935, "learning_rate": 0.0002, "epoch": 0.03084495488105004, "step": 470}, {"loss": 1.0685, "grad_norm": 0.2637856900691986, "learning_rate": 0.0002, "epoch": 0.031501230516817066, "step": 480}, {"loss": 0.9879, "grad_norm": 0.3064589500427246, "learning_rate": 0.0002, "epoch": 0.03215750615258409, "step": 490}, {"loss": 0.9657, "grad_norm": 0.25345391035079956, "learning_rate": 0.0002, "epoch": 0.03281378178835111, "step": 500}, {"loss": 0.9518, "grad_norm": 0.3100789785385132, "learning_rate": 0.0002, "epoch": 0.03347005742411813, "step": 510}, {"loss": 1.0108, "grad_norm": 0.3312002420425415, "learning_rate": 0.0002, "epoch": 0.034126333059885154, "step": 520}, {"loss": 1.0157, "grad_norm": 0.23432421684265137, "learning_rate": 0.0002, "epoch": 0.034782608695652174, "step": 530}, {"loss": 1.0093, "grad_norm": 0.3079119622707367, "learning_rate": 0.0002, "epoch": 0.035438884331419195, "step": 540}, {"loss": 0.9534, "grad_norm": 0.2555035650730133, "learning_rate": 0.0002, "epoch": 0.03609515996718622, "step": 550}, {"loss": 0.9648, "grad_norm": 0.3530837893486023, "learning_rate": 0.0002, "epoch": 0.03675143560295324, "step": 560}, {"loss": 1.0432, "grad_norm": 0.26614823937416077, "learning_rate": 0.0002, "epoch": 0.03740771123872026, "step": 570}, {"loss": 1.0108, "grad_norm": 0.2618412971496582, "learning_rate": 0.0002, "epoch": 0.03806398687448728, "step": 580}, {"loss": 0.9834, "grad_norm": 0.26110127568244934, "learning_rate": 0.0002, "epoch": 0.03872026251025431, "step": 590}, {"loss": 0.9766, "grad_norm": 0.3050612211227417, "learning_rate": 0.0002, "epoch": 0.03937653814602133, "step": 600}, {"loss": 1.0195, "grad_norm": 0.27394089102745056, "learning_rate": 0.0002, "epoch": 0.04003281378178835, "step": 610}, {"loss": 0.9776, "grad_norm": 0.25288277864456177, "learning_rate": 0.0002, "epoch": 0.04068908941755537, "step": 620}, {"loss": 0.9199, "grad_norm": 0.2835882902145386, "learning_rate": 0.0002, "epoch": 0.0413453650533224, "step": 630}, {"loss": 0.9605, "grad_norm": 0.2621902823448181, "learning_rate": 0.0002, "epoch": 0.04200164068908942, "step": 640}, {"loss": 0.9494, "grad_norm": 0.23218439519405365, "learning_rate": 0.0002, "epoch": 0.04265791632485644, "step": 650}, {"loss": 0.994, "grad_norm": 0.26911118626594543, "learning_rate": 0.0002, "epoch": 0.043314191960623465, "step": 660}, {"loss": 0.972, "grad_norm": 0.25920751690864563, "learning_rate": 0.0002, "epoch": 0.043970467596390485, "step": 670}, {"loss": 0.9774, "grad_norm": 0.2772065997123718, "learning_rate": 0.0002, "epoch": 0.044626743232157505, "step": 680}, {"loss": 0.9114, "grad_norm": 0.275421142578125, "learning_rate": 0.0002, "epoch": 0.045283018867924525, "step": 690}, {"loss": 1.028, "grad_norm": 0.23931777477264404, "learning_rate": 0.0002, "epoch": 0.04593929450369155, "step": 700}, {"loss": 0.9309, "grad_norm": 0.3031066060066223, "learning_rate": 0.0002, "epoch": 0.04659557013945857, "step": 710}, {"loss": 0.9511, "grad_norm": 0.2655068039894104, "learning_rate": 0.0002, "epoch": 0.04725184577522559, "step": 720}, {"loss": 0.9012, "grad_norm": 0.26064610481262207, "learning_rate": 0.0002, "epoch": 0.04790812141099262, "step": 730}, {"loss": 0.9809, "grad_norm": 0.2934698760509491, "learning_rate": 0.0002, "epoch": 0.04856439704675964, "step": 740}, {"loss": 0.8859, "grad_norm": 0.2672717571258545, "learning_rate": 0.0002, "epoch": 0.04922067268252666, "step": 750}, {"loss": 1.0251, "grad_norm": 0.27246803045272827, "learning_rate": 0.0002, "epoch": 0.04987694831829368, "step": 760}, {"loss": 0.977, "grad_norm": 0.27560853958129883, "learning_rate": 0.0002, "epoch": 0.05053322395406071, "step": 770}, {"loss": 0.9735, "grad_norm": 0.2449599653482437, "learning_rate": 0.0002, "epoch": 0.05118949958982773, "step": 780}, {"loss": 0.9765, "grad_norm": 0.2873939871788025, "learning_rate": 0.0002, "epoch": 0.05184577522559475, "step": 790}, {"loss": 0.9647, "grad_norm": 0.2875595688819885, "learning_rate": 0.0002, "epoch": 0.052502050861361775, "step": 800}, {"loss": 1.0379, "grad_norm": 0.4079909324645996, "learning_rate": 0.0002, "epoch": 0.053158326497128795, "step": 810}, {"loss": 0.9782, "grad_norm": 0.2733079195022583, "learning_rate": 0.0002, "epoch": 0.053814602132895815, "step": 820}, {"loss": 0.98, "grad_norm": 0.2624184489250183, "learning_rate": 0.0002, "epoch": 0.054470877768662836, "step": 830}, {"loss": 1.0212, "grad_norm": 0.279851496219635, "learning_rate": 0.0002, "epoch": 0.05512715340442986, "step": 840}, {"loss": 0.8994, "grad_norm": 0.2573111951351166, "learning_rate": 0.0002, "epoch": 0.05578342904019688, "step": 850}, {"loss": 0.9823, "grad_norm": 0.28704535961151123, "learning_rate": 0.0002, "epoch": 0.0564397046759639, "step": 860}, {"loss": 0.9744, "grad_norm": 0.2664150297641754, "learning_rate": 0.0002, "epoch": 0.05709598031173093, "step": 870}, {"loss": 0.9379, "grad_norm": 0.2858041822910309, "learning_rate": 0.0002, "epoch": 0.05775225594749795, "step": 880}, {"loss": 0.9453, "grad_norm": 0.2394150048494339, "learning_rate": 0.0002, "epoch": 0.05840853158326497, "step": 890}, {"loss": 0.9372, "grad_norm": 0.30714845657348633, "learning_rate": 0.0002, "epoch": 0.05906480721903199, "step": 900}, {"loss": 0.9677, "grad_norm": 0.2740330994129181, "learning_rate": 0.0002, "epoch": 0.05972108285479902, "step": 910}, {"loss": 1.0112, "grad_norm": 0.33505478501319885, "learning_rate": 0.0002, "epoch": 0.06037735849056604, "step": 920}, {"loss": 0.986, "grad_norm": 0.2945438325405121, "learning_rate": 0.0002, "epoch": 0.06103363412633306, "step": 930}, {"loss": 0.9498, "grad_norm": 0.24749146401882172, "learning_rate": 0.0002, "epoch": 0.06168990976210008, "step": 940}, {"loss": 1.0357, "grad_norm": 0.3315669894218445, "learning_rate": 0.0002, "epoch": 0.062346185397867106, "step": 950}, {"loss": 0.9698, "grad_norm": 0.27253520488739014, "learning_rate": 0.0002, "epoch": 0.06300246103363413, "step": 960}, {"loss": 0.9544, "grad_norm": 0.26884031295776367, "learning_rate": 0.0002, "epoch": 0.06365873666940115, "step": 970}, {"loss": 0.9697, "grad_norm": 0.286920964717865, "learning_rate": 0.0002, "epoch": 0.06431501230516817, "step": 980}, {"loss": 0.9704, "grad_norm": 0.28334400057792664, "learning_rate": 0.0002, "epoch": 0.0649712879409352, "step": 990}, {"loss": 0.917, "grad_norm": 0.2672102749347687, "learning_rate": 0.0002, "epoch": 0.06562756357670221, "step": 1000}, {"loss": 1.0062, "grad_norm": 0.3247123062610626, "learning_rate": 0.0002, "epoch": 0.06628383921246923, "step": 1010}, {"loss": 0.9716, "grad_norm": 0.259440541267395, "learning_rate": 0.0002, "epoch": 0.06694011484823625, "step": 1020}, {"loss": 0.9569, "grad_norm": 0.2795625329017639, "learning_rate": 0.0002, "epoch": 0.06759639048400329, "step": 1030}, {"loss": 0.9842, "grad_norm": 0.2784935534000397, "learning_rate": 0.0002, "epoch": 0.06825266611977031, "step": 1040}, {"loss": 0.9171, "grad_norm": 0.24605989456176758, "learning_rate": 0.0002, "epoch": 0.06890894175553733, "step": 1050}, {"loss": 0.9538, "grad_norm": 0.25421491265296936, "learning_rate": 0.0002, "epoch": 0.06956521739130435, "step": 1060}, {"loss": 1.0142, "grad_norm": 0.2693536579608917, "learning_rate": 0.0002, "epoch": 0.07022149302707137, "step": 1070}, {"loss": 0.9436, "grad_norm": 0.28166458010673523, "learning_rate": 0.0002, "epoch": 0.07087776866283839, "step": 1080}, {"loss": 0.9187, "grad_norm": 0.2752484679222107, "learning_rate": 0.0002, "epoch": 0.07153404429860541, "step": 1090}, {"loss": 1.0069, "grad_norm": 0.24141381680965424, "learning_rate": 0.0002, "epoch": 0.07219031993437244, "step": 1100}, {"loss": 0.9282, "grad_norm": 0.2347770482301712, "learning_rate": 0.0002, "epoch": 0.07284659557013946, "step": 1110}, {"loss": 0.978, "grad_norm": 0.29999610781669617, "learning_rate": 0.0002, "epoch": 0.07350287120590648, "step": 1120}, {"loss": 0.9763, "grad_norm": 0.2811068892478943, "learning_rate": 0.0002, "epoch": 0.0741591468416735, "step": 1130}, {"loss": 0.978, "grad_norm": 0.24790801107883453, "learning_rate": 0.0002, "epoch": 0.07481542247744052, "step": 1140}, {"loss": 1.001, "grad_norm": 0.3251914978027344, "learning_rate": 0.0002, "epoch": 0.07547169811320754, "step": 1150}, {"loss": 1.0407, "grad_norm": 0.2616347074508667, "learning_rate": 0.0002, "epoch": 0.07612797374897456, "step": 1160}, {"loss": 0.978, "grad_norm": 0.2568797171115875, "learning_rate": 0.0002, "epoch": 0.07678424938474158, "step": 1170}, {"loss": 0.9851, "grad_norm": 0.2693248689174652, "learning_rate": 0.0002, "epoch": 0.07744052502050862, "step": 1180}, {"loss": 0.996, "grad_norm": 0.3270677924156189, "learning_rate": 0.0002, "epoch": 0.07809680065627564, "step": 1190}, {"loss": 0.9751, "grad_norm": 0.2481861710548401, "learning_rate": 0.0002, "epoch": 0.07875307629204266, "step": 1200}, {"loss": 1.0047, "grad_norm": 0.2612398564815521, "learning_rate": 0.0002, "epoch": 0.07940935192780968, "step": 1210}, {"loss": 0.9789, "grad_norm": 0.26566916704177856, "learning_rate": 0.0002, "epoch": 0.0800656275635767, "step": 1220}, {"loss": 1.0468, "grad_norm": 0.28026407957077026, "learning_rate": 0.0002, "epoch": 0.08072190319934372, "step": 1230}, {"loss": 0.9651, "grad_norm": 0.25609225034713745, "learning_rate": 0.0002, "epoch": 0.08137817883511074, "step": 1240}, {"loss": 0.9648, "grad_norm": 0.29063138365745544, "learning_rate": 0.0002, "epoch": 0.08203445447087777, "step": 1250}, {"loss": 0.9302, "grad_norm": 0.25998231768608093, "learning_rate": 0.0002, "epoch": 0.0826907301066448, "step": 1260}, {"loss": 0.9733, "grad_norm": 0.295261412858963, "learning_rate": 0.0002, "epoch": 0.08334700574241181, "step": 1270}, {"loss": 0.9775, "grad_norm": 0.2479529082775116, "learning_rate": 0.0002, "epoch": 0.08400328137817883, "step": 1280}, {"loss": 0.9891, "grad_norm": 0.25226080417633057, "learning_rate": 0.0002, "epoch": 0.08465955701394585, "step": 1290}, {"loss": 0.9693, "grad_norm": 0.290462851524353, "learning_rate": 0.0002, "epoch": 0.08531583264971287, "step": 1300}, {"loss": 1.0319, "grad_norm": 0.2832583785057068, "learning_rate": 0.0002, "epoch": 0.0859721082854799, "step": 1310}, {"loss": 1.0105, "grad_norm": 0.2515616714954376, "learning_rate": 0.0002, "epoch": 0.08662838392124693, "step": 1320}, {"loss": 0.9717, "grad_norm": 0.28894907236099243, "learning_rate": 0.0002, "epoch": 0.08728465955701395, "step": 1330}, {"loss": 0.9736, "grad_norm": 0.2881310284137726, "learning_rate": 0.0002, "epoch": 0.08794093519278097, "step": 1340}, {"loss": 1.0107, "grad_norm": 0.2654068171977997, "learning_rate": 0.0002, "epoch": 0.08859721082854799, "step": 1350}, {"loss": 1.0249, "grad_norm": 0.2929916977882385, "learning_rate": 0.0002, "epoch": 0.08925348646431501, "step": 1360}, {"loss": 0.9605, "grad_norm": 0.27295321226119995, "learning_rate": 0.0002, "epoch": 0.08990976210008203, "step": 1370}, {"loss": 0.9709, "grad_norm": 0.27720171213150024, "learning_rate": 0.0002, "epoch": 0.09056603773584905, "step": 1380}, {"loss": 1.0449, "grad_norm": 0.2784966826438904, "learning_rate": 0.0002, "epoch": 0.09122231337161608, "step": 1390}, {"loss": 0.9442, "grad_norm": 0.25110408663749695, "learning_rate": 0.0002, "epoch": 0.0918785890073831, "step": 1400}, {"loss": 0.9552, "grad_norm": 0.2501158118247986, "learning_rate": 0.0002, "epoch": 0.09253486464315012, "step": 1410}, {"loss": 0.9493, "grad_norm": 0.3097717761993408, "learning_rate": 0.0002, "epoch": 0.09319114027891715, "step": 1420}, {"loss": 0.9635, "grad_norm": 0.257308691740036, "learning_rate": 0.0002, "epoch": 0.09384741591468417, "step": 1430}, {"loss": 0.9749, "grad_norm": 0.2885759472846985, "learning_rate": 0.0002, "epoch": 0.09450369155045119, "step": 1440}, {"loss": 0.9592, "grad_norm": 0.2938264012336731, "learning_rate": 0.0002, "epoch": 0.0951599671862182, "step": 1450}, {"loss": 1.0326, "grad_norm": 0.2543550729751587, "learning_rate": 0.0002, "epoch": 0.09581624282198524, "step": 1460}, {"loss": 0.9433, "grad_norm": 0.3020402789115906, "learning_rate": 0.0002, "epoch": 0.09647251845775226, "step": 1470}, {"loss": 0.9132, "grad_norm": 0.2818080186843872, "learning_rate": 0.0002, "epoch": 0.09712879409351928, "step": 1480}, {"loss": 0.9217, "grad_norm": 0.2420656979084015, "learning_rate": 0.0002, "epoch": 0.0977850697292863, "step": 1490}, {"loss": 0.938, "grad_norm": 0.31427690386772156, "learning_rate": 0.0002, "epoch": 0.09844134536505332, "step": 1500}, {"loss": 0.9234, "grad_norm": 0.27618250250816345, "learning_rate": 0.0002, "epoch": 0.09909762100082034, "step": 1510}, {"loss": 0.9019, "grad_norm": 0.26401254534721375, "learning_rate": 0.0002, "epoch": 0.09975389663658736, "step": 1520}, {"loss": 0.9992, "grad_norm": 0.23575739562511444, "learning_rate": 0.0002, "epoch": 0.1004101722723544, "step": 1530}, {"loss": 0.9315, "grad_norm": 0.3076087534427643, "learning_rate": 0.0002, "epoch": 0.10106644790812142, "step": 1540}, {"loss": 1.0138, "grad_norm": 0.28179317712783813, "learning_rate": 0.0002, "epoch": 0.10172272354388844, "step": 1550}, {"loss": 0.9587, "grad_norm": 0.28638190031051636, "learning_rate": 0.0002, "epoch": 0.10237899917965546, "step": 1560}, {"loss": 0.9576, "grad_norm": 0.24788478016853333, "learning_rate": 0.0002, "epoch": 0.10303527481542248, "step": 1570}, {"loss": 0.9523, "grad_norm": 0.2987830340862274, "learning_rate": 0.0002, "epoch": 0.1036915504511895, "step": 1580}, {"loss": 0.9782, "grad_norm": 0.3252484202384949, "learning_rate": 0.0002, "epoch": 0.10434782608695652, "step": 1590}, {"loss": 0.9378, "grad_norm": 0.2763068377971649, "learning_rate": 0.0002, "epoch": 0.10500410172272355, "step": 1600}, {"loss": 0.9491, "grad_norm": 0.3219335973262787, "learning_rate": 0.0002, "epoch": 0.10566037735849057, "step": 1610}, {"loss": 0.9181, "grad_norm": 0.2708939015865326, "learning_rate": 0.0002, "epoch": 0.10631665299425759, "step": 1620}, {"loss": 1.0008, "grad_norm": 0.29320818185806274, "learning_rate": 0.0002, "epoch": 0.10697292863002461, "step": 1630}, {"loss": 0.9593, "grad_norm": 0.2731851041316986, "learning_rate": 0.0002, "epoch": 0.10762920426579163, "step": 1640}, {"loss": 0.9915, "grad_norm": 0.29563331604003906, "learning_rate": 0.0002, "epoch": 0.10828547990155865, "step": 1650}, {"loss": 0.9543, "grad_norm": 0.33387669920921326, "learning_rate": 0.0002, "epoch": 0.10894175553732567, "step": 1660}, {"loss": 0.941, "grad_norm": 0.3157867193222046, "learning_rate": 0.0002, "epoch": 0.1095980311730927, "step": 1670}, {"loss": 0.9107, "grad_norm": 0.24688409268856049, "learning_rate": 0.0002, "epoch": 0.11025430680885973, "step": 1680}, {"loss": 0.9392, "grad_norm": 0.2868671417236328, "learning_rate": 0.0002, "epoch": 0.11091058244462675, "step": 1690}, {"loss": 0.9476, "grad_norm": 0.2555101811885834, "learning_rate": 0.0002, "epoch": 0.11156685808039377, "step": 1700}, {"loss": 0.9261, "grad_norm": 0.3224661946296692, "learning_rate": 0.0002, "epoch": 0.11222313371616079, "step": 1710}, {"loss": 0.9372, "grad_norm": 0.28927505016326904, "learning_rate": 0.0002, "epoch": 0.1128794093519278, "step": 1720}, {"loss": 0.9237, "grad_norm": 0.25575608015060425, "learning_rate": 0.0002, "epoch": 0.11353568498769483, "step": 1730}, {"loss": 0.9526, "grad_norm": 0.27490100264549255, "learning_rate": 0.0002, "epoch": 0.11419196062346186, "step": 1740}, {"loss": 1.0311, "grad_norm": 0.28235989809036255, "learning_rate": 0.0002, "epoch": 0.11484823625922888, "step": 1750}, {"loss": 0.9112, "grad_norm": 0.2636060416698456, "learning_rate": 0.0002, "epoch": 0.1155045118949959, "step": 1760}, {"loss": 0.9249, "grad_norm": 0.2642059624195099, "learning_rate": 0.0002, "epoch": 0.11616078753076292, "step": 1770}, {"loss": 0.964, "grad_norm": 0.261807918548584, "learning_rate": 0.0002, "epoch": 0.11681706316652994, "step": 1780}, {"loss": 0.9473, "grad_norm": 0.24618281424045563, "learning_rate": 0.0002, "epoch": 0.11747333880229696, "step": 1790}, {"loss": 0.9452, "grad_norm": 0.2700425982475281, "learning_rate": 0.0002, "epoch": 0.11812961443806398, "step": 1800}, {"loss": 0.9687, "grad_norm": 0.22687454521656036, "learning_rate": 0.0002, "epoch": 0.11878589007383102, "step": 1810}, {"loss": 0.9099, "grad_norm": 0.2843499481678009, "learning_rate": 0.0002, "epoch": 0.11944216570959804, "step": 1820}, {"loss": 0.9295, "grad_norm": 0.23511677980422974, "learning_rate": 0.0002, "epoch": 0.12009844134536506, "step": 1830}, {"loss": 0.998, "grad_norm": 0.34054702520370483, "learning_rate": 0.0002, "epoch": 0.12075471698113208, "step": 1840}, {"loss": 0.9343, "grad_norm": 0.32845374941825867, "learning_rate": 0.0002, "epoch": 0.1214109926168991, "step": 1850}, {"loss": 0.983, "grad_norm": 0.2843034267425537, "learning_rate": 0.0002, "epoch": 0.12206726825266612, "step": 1860}, {"loss": 0.963, "grad_norm": 0.2824552357196808, "learning_rate": 0.0002, "epoch": 0.12272354388843314, "step": 1870}, {"loss": 0.9875, "grad_norm": 0.26878267526626587, "learning_rate": 0.0002, "epoch": 0.12337981952420016, "step": 1880}, {"loss": 0.9784, "grad_norm": 0.31207871437072754, "learning_rate": 0.0002, "epoch": 0.12403609515996719, "step": 1890}, {"loss": 0.9497, "grad_norm": 0.2617819905281067, "learning_rate": 0.0002, "epoch": 0.12469237079573421, "step": 1900}, {"loss": 0.8948, "grad_norm": 0.2652885615825653, "learning_rate": 0.0002, "epoch": 0.12534864643150123, "step": 1910}, {"loss": 0.9534, "grad_norm": 0.29473352432250977, "learning_rate": 0.0002, "epoch": 0.12600492206726827, "step": 1920}, {"loss": 0.9695, "grad_norm": 0.3101664185523987, "learning_rate": 0.0002, "epoch": 0.12666119770303527, "step": 1930}, {"loss": 0.9928, "grad_norm": 0.2714068293571472, "learning_rate": 0.0002, "epoch": 0.1273174733388023, "step": 1940}, {"loss": 1.0293, "grad_norm": 0.2684655487537384, "learning_rate": 0.0002, "epoch": 0.1279737489745693, "step": 1950}, {"loss": 0.9236, "grad_norm": 0.2543509900569916, "learning_rate": 0.0002, "epoch": 0.12863002461033635, "step": 1960}, {"loss": 0.9598, "grad_norm": 0.25268790125846863, "learning_rate": 0.0002, "epoch": 0.12928630024610335, "step": 1970}, {"loss": 0.967, "grad_norm": 0.27713078260421753, "learning_rate": 0.0002, "epoch": 0.1299425758818704, "step": 1980}, {"loss": 0.9411, "grad_norm": 0.3185126483440399, "learning_rate": 0.0002, "epoch": 0.13059885151763742, "step": 1990}, {"loss": 0.9864, "grad_norm": 0.26718857884407043, "learning_rate": 0.0002, "epoch": 0.13125512715340443, "step": 2000}, {"loss": 0.9728, "grad_norm": 0.2843841314315796, "learning_rate": 0.0002, "epoch": 0.13191140278917146, "step": 2010}, {"loss": 0.993, "grad_norm": 0.3013055622577667, "learning_rate": 0.0002, "epoch": 0.13256767842493847, "step": 2020}, {"loss": 0.9437, "grad_norm": 0.27985867857933044, "learning_rate": 0.0002, "epoch": 0.1332239540607055, "step": 2030}, {"loss": 0.905, "grad_norm": 0.2501908242702484, "learning_rate": 0.0002, "epoch": 0.1338802296964725, "step": 2040}, {"loss": 0.9603, "grad_norm": 0.441340833902359, "learning_rate": 0.0002, "epoch": 0.13453650533223954, "step": 2050}, {"loss": 0.9166, "grad_norm": 2.568060874938965, "learning_rate": 0.0002, "epoch": 0.13519278096800658, "step": 2060}, {"loss": 0.9803, "grad_norm": 0.2505454421043396, "learning_rate": 0.0002, "epoch": 0.13584905660377358, "step": 2070}, {"loss": 1.0159, "grad_norm": 0.33997446298599243, "learning_rate": 0.0002, "epoch": 0.13650533223954062, "step": 2080}, {"loss": 0.908, "grad_norm": 0.28161343932151794, "learning_rate": 0.0002, "epoch": 0.13716160787530762, "step": 2090}, {"loss": 0.9709, "grad_norm": 0.28142687678337097, "learning_rate": 0.0002, "epoch": 0.13781788351107466, "step": 2100}, {"loss": 0.9716, "grad_norm": 0.3137170076370239, "learning_rate": 0.0002, "epoch": 0.13847415914684166, "step": 2110}, {"loss": 0.9698, "grad_norm": 0.3072240352630615, "learning_rate": 0.0002, "epoch": 0.1391304347826087, "step": 2120}, {"loss": 0.9979, "grad_norm": 0.24328380823135376, "learning_rate": 0.0002, "epoch": 0.13978671041837573, "step": 2130}, {"loss": 0.9549, "grad_norm": 0.3065047860145569, "learning_rate": 0.0002, "epoch": 0.14044298605414274, "step": 2140}, {"loss": 0.9897, "grad_norm": 0.34212175011634827, "learning_rate": 0.0002, "epoch": 0.14109926168990977, "step": 2150}, {"loss": 0.9601, "grad_norm": 0.27491796016693115, "learning_rate": 0.0002, "epoch": 0.14175553732567678, "step": 2160}, {"loss": 0.9928, "grad_norm": 0.26518693566322327, "learning_rate": 0.0002, "epoch": 0.1424118129614438, "step": 2170}, {"loss": 0.9385, "grad_norm": 0.28350934386253357, "learning_rate": 0.0002, "epoch": 0.14306808859721082, "step": 2180}, {"loss": 0.9878, "grad_norm": 0.3287768065929413, "learning_rate": 0.0002, "epoch": 0.14372436423297785, "step": 2190}, {"loss": 0.9192, "grad_norm": 0.26362666487693787, "learning_rate": 0.0002, "epoch": 0.1443806398687449, "step": 2200}, {"loss": 0.9804, "grad_norm": 0.31169235706329346, "learning_rate": 0.0002, "epoch": 0.1450369155045119, "step": 2210}, {"loss": 0.9857, "grad_norm": 0.258667528629303, "learning_rate": 0.0002, "epoch": 0.14569319114027893, "step": 2220}, {"loss": 0.9898, "grad_norm": 0.32289111614227295, "learning_rate": 0.0002, "epoch": 0.14634946677604593, "step": 2230}, {"loss": 1.0051, "grad_norm": 0.3344270586967468, "learning_rate": 0.0002, "epoch": 0.14700574241181297, "step": 2240}, {"loss": 0.9337, "grad_norm": 0.3001033067703247, "learning_rate": 0.0002, "epoch": 0.14766201804757997, "step": 2250}, {"loss": 0.9788, "grad_norm": 0.2718261778354645, "learning_rate": 0.0002, "epoch": 0.148318293683347, "step": 2260}, {"loss": 0.9801, "grad_norm": 0.3059164583683014, "learning_rate": 0.0002, "epoch": 0.14897456931911401, "step": 2270}, {"loss": 0.8914, "grad_norm": 0.2939850389957428, "learning_rate": 0.0002, "epoch": 0.14963084495488105, "step": 2280}, {"loss": 0.951, "grad_norm": 0.2803564965724945, "learning_rate": 0.0002, "epoch": 0.15028712059064808, "step": 2290}, {"loss": 0.9131, "grad_norm": 0.2843068242073059, "learning_rate": 0.0002, "epoch": 0.1509433962264151, "step": 2300}, {"loss": 0.9721, "grad_norm": 0.28517085313796997, "learning_rate": 0.0002, "epoch": 0.15159967186218212, "step": 2310}, {"loss": 0.9402, "grad_norm": 0.291030615568161, "learning_rate": 0.0002, "epoch": 0.15225594749794913, "step": 2320}, {"loss": 0.9509, "grad_norm": 0.2712008059024811, "learning_rate": 0.0002, "epoch": 0.15291222313371616, "step": 2330}, {"loss": 0.9406, "grad_norm": 0.3357657790184021, "learning_rate": 0.0002, "epoch": 0.15356849876948317, "step": 2340}, {"loss": 0.961, "grad_norm": 0.28412291407585144, "learning_rate": 0.0002, "epoch": 0.1542247744052502, "step": 2350}, {"loss": 0.9254, "grad_norm": 0.3217862844467163, "learning_rate": 0.0002, "epoch": 0.15488105004101724, "step": 2360}, {"loss": 0.9172, "grad_norm": 0.32076528668403625, "learning_rate": 0.0002, "epoch": 0.15553732567678424, "step": 2370}, {"loss": 0.9459, "grad_norm": 0.3490257263183594, "learning_rate": 0.0002, "epoch": 0.15619360131255128, "step": 2380}, {"loss": 0.9051, "grad_norm": 0.28253331780433655, "learning_rate": 0.0002, "epoch": 0.15684987694831828, "step": 2390}, {"loss": 0.9597, "grad_norm": 0.2875654399394989, "learning_rate": 0.0002, "epoch": 0.15750615258408532, "step": 2400}, {"loss": 0.9164, "grad_norm": 0.3347921669483185, "learning_rate": 0.0002, "epoch": 0.15816242821985232, "step": 2410}, {"loss": 1.0023, "grad_norm": 0.2624322474002838, "learning_rate": 0.0002, "epoch": 0.15881870385561936, "step": 2420}, {"loss": 0.9751, "grad_norm": 0.29683780670166016, "learning_rate": 0.0002, "epoch": 0.1594749794913864, "step": 2430}, {"loss": 0.967, "grad_norm": 0.29237303137779236, "learning_rate": 0.0002, "epoch": 0.1601312551271534, "step": 2440}, {"loss": 0.9754, "grad_norm": 0.3541257977485657, "learning_rate": 0.0002, "epoch": 0.16078753076292043, "step": 2450}, {"loss": 0.8977, "grad_norm": 0.28203415870666504, "learning_rate": 0.0002, "epoch": 0.16144380639868744, "step": 2460}, {"loss": 0.9489, "grad_norm": 0.322329044342041, "learning_rate": 0.0002, "epoch": 0.16210008203445447, "step": 2470}, {"loss": 0.9224, "grad_norm": 0.27731558680534363, "learning_rate": 0.0002, "epoch": 0.16275635767022148, "step": 2480}, {"loss": 0.9773, "grad_norm": 0.30348825454711914, "learning_rate": 0.0002, "epoch": 0.16341263330598851, "step": 2490}, {"loss": 0.9318, "grad_norm": 0.29504773020744324, "learning_rate": 0.0002, "epoch": 0.16406890894175555, "step": 2500}, {"loss": 0.8931, "grad_norm": 0.2977028489112854, "learning_rate": 0.0002, "epoch": 0.16472518457752255, "step": 2510}, {"loss": 0.953, "grad_norm": 0.3172457218170166, "learning_rate": 0.0002, "epoch": 0.1653814602132896, "step": 2520}, {"loss": 0.9786, "grad_norm": 0.3320612907409668, "learning_rate": 0.0002, "epoch": 0.1660377358490566, "step": 2530}, {"loss": 0.895, "grad_norm": 0.2473023235797882, "learning_rate": 0.0002, "epoch": 0.16669401148482363, "step": 2540}, {"loss": 0.9692, "grad_norm": 0.3250006139278412, "learning_rate": 0.0002, "epoch": 0.16735028712059064, "step": 2550}, {"loss": 0.956, "grad_norm": 0.2940689027309418, "learning_rate": 0.0002, "epoch": 0.16800656275635767, "step": 2560}, {"loss": 0.8926, "grad_norm": 0.34614887833595276, "learning_rate": 0.0002, "epoch": 0.1686628383921247, "step": 2570}, {"loss": 1.0153, "grad_norm": 0.3513164520263672, "learning_rate": 0.0002, "epoch": 0.1693191140278917, "step": 2580}, {"loss": 0.9608, "grad_norm": 0.2530531585216522, "learning_rate": 0.0002, "epoch": 0.16997538966365874, "step": 2590}, {"loss": 0.9876, "grad_norm": 0.3028896152973175, "learning_rate": 0.0002, "epoch": 0.17063166529942575, "step": 2600}, {"loss": 0.9415, "grad_norm": 0.304739773273468, "learning_rate": 0.0002, "epoch": 0.17128794093519278, "step": 2610}, {"loss": 0.9857, "grad_norm": 0.3360660672187805, "learning_rate": 0.0002, "epoch": 0.1719442165709598, "step": 2620}, {"loss": 0.9173, "grad_norm": 0.24901753664016724, "learning_rate": 0.0002, "epoch": 0.17260049220672682, "step": 2630}, {"loss": 0.9351, "grad_norm": 0.32212400436401367, "learning_rate": 0.0002, "epoch": 0.17325676784249386, "step": 2640}, {"loss": 0.9284, "grad_norm": 0.2780437767505646, "learning_rate": 0.0002, "epoch": 0.17391304347826086, "step": 2650}, {"loss": 0.9368, "grad_norm": 0.25902262330055237, "learning_rate": 0.0002, "epoch": 0.1745693191140279, "step": 2660}, {"loss": 0.9378, "grad_norm": 0.29171422123908997, "learning_rate": 0.0002, "epoch": 0.1752255947497949, "step": 2670}, {"loss": 0.9525, "grad_norm": 0.2586783766746521, "learning_rate": 0.0002, "epoch": 0.17588187038556194, "step": 2680}, {"loss": 0.9367, "grad_norm": 0.25609949231147766, "learning_rate": 0.0002, "epoch": 0.17653814602132895, "step": 2690}, {"loss": 0.8675, "grad_norm": 0.29789483547210693, "learning_rate": 0.0002, "epoch": 0.17719442165709598, "step": 2700}, {"loss": 0.9918, "grad_norm": 0.2938411235809326, "learning_rate": 0.0002, "epoch": 0.177850697292863, "step": 2710}, {"loss": 0.9824, "grad_norm": 0.31436532735824585, "learning_rate": 0.0002, "epoch": 0.17850697292863002, "step": 2720}, {"loss": 1.0094, "grad_norm": 0.31310203671455383, "learning_rate": 0.0002, "epoch": 0.17916324856439705, "step": 2730}, {"loss": 0.9105, "grad_norm": 0.2695367634296417, "learning_rate": 0.0002, "epoch": 0.17981952420016406, "step": 2740}, {"loss": 0.9216, "grad_norm": 0.32590436935424805, "learning_rate": 0.0002, "epoch": 0.1804757998359311, "step": 2750}, {"loss": 0.923, "grad_norm": 0.32683756947517395, "learning_rate": 0.0002, "epoch": 0.1811320754716981, "step": 2760}, {"loss": 0.938, "grad_norm": 0.33043643832206726, "learning_rate": 0.0002, "epoch": 0.18178835110746513, "step": 2770}, {"loss": 0.937, "grad_norm": 0.3137816786766052, "learning_rate": 0.0002, "epoch": 0.18244462674323217, "step": 2780}, {"loss": 0.8923, "grad_norm": 0.33037737011909485, "learning_rate": 0.0002, "epoch": 0.18310090237899918, "step": 2790}, {"loss": 0.9179, "grad_norm": 0.27679184079170227, "learning_rate": 0.0002, "epoch": 0.1837571780147662, "step": 2800}, {"loss": 0.9047, "grad_norm": 0.42975902557373047, "learning_rate": 0.0002, "epoch": 0.18441345365053322, "step": 2810}, {"loss": 0.9431, "grad_norm": 0.2915041744709015, "learning_rate": 0.0002, "epoch": 0.18506972928630025, "step": 2820}, {"loss": 0.9573, "grad_norm": 0.315197229385376, "learning_rate": 0.0002, "epoch": 0.18572600492206726, "step": 2830}, {"loss": 0.9909, "grad_norm": 0.31129002571105957, "learning_rate": 0.0002, "epoch": 0.1863822805578343, "step": 2840}, {"loss": 0.9222, "grad_norm": 0.2783232033252716, "learning_rate": 0.0002, "epoch": 0.18703855619360132, "step": 2850}, {"loss": 0.9347, "grad_norm": 0.34501492977142334, "learning_rate": 0.0002, "epoch": 0.18769483182936833, "step": 2860}, {"loss": 0.949, "grad_norm": 0.32616767287254333, "learning_rate": 0.0002, "epoch": 0.18835110746513536, "step": 2870}, {"loss": 0.9909, "grad_norm": 0.25213682651519775, "learning_rate": 0.0002, "epoch": 0.18900738310090237, "step": 2880}, {"loss": 0.892, "grad_norm": 0.2745118737220764, "learning_rate": 0.0002, "epoch": 0.1896636587366694, "step": 2890}, {"loss": 0.9737, "grad_norm": 0.26175656914711, "learning_rate": 0.0002, "epoch": 0.1903199343724364, "step": 2900}, {"loss": 0.8853, "grad_norm": 0.3019673228263855, "learning_rate": 0.0002, "epoch": 0.19097621000820345, "step": 2910}, {"loss": 1.0016, "grad_norm": 0.30918899178504944, "learning_rate": 0.0002, "epoch": 0.19163248564397048, "step": 2920}, {"loss": 0.9301, "grad_norm": 0.320893794298172, "learning_rate": 0.0002, "epoch": 0.19228876127973749, "step": 2930}, {"loss": 0.9592, "grad_norm": 0.2736368477344513, "learning_rate": 0.0002, "epoch": 0.19294503691550452, "step": 2940}, {"loss": 0.9334, "grad_norm": 0.28817251324653625, "learning_rate": 0.0002, "epoch": 0.19360131255127153, "step": 2950}, {"loss": 0.9484, "grad_norm": 0.26863837242126465, "learning_rate": 0.0002, "epoch": 0.19425758818703856, "step": 2960}, {"loss": 0.9466, "grad_norm": 0.28012585639953613, "learning_rate": 0.0002, "epoch": 0.19491386382280557, "step": 2970}, {"loss": 0.9623, "grad_norm": 0.2929932773113251, "learning_rate": 0.0002, "epoch": 0.1955701394585726, "step": 2980}, {"loss": 0.9406, "grad_norm": 0.29878300428390503, "learning_rate": 0.0002, "epoch": 0.19622641509433963, "step": 2990}, {"loss": 0.9714, "grad_norm": 0.2573733329772949, "learning_rate": 0.0002, "epoch": 0.19688269073010664, "step": 3000}, {"loss": 0.9763, "grad_norm": 0.31736820936203003, "learning_rate": 0.0002, "epoch": 0.19753896636587367, "step": 3010}, {"loss": 1.0111, "grad_norm": 0.2864682972431183, "learning_rate": 0.0002, "epoch": 0.19819524200164068, "step": 3020}, {"loss": 0.9193, "grad_norm": 0.3161790668964386, "learning_rate": 0.0002, "epoch": 0.19885151763740772, "step": 3030}, {"loss": 0.9298, "grad_norm": 0.37538009881973267, "learning_rate": 0.0002, "epoch": 0.19950779327317472, "step": 3040}, {"loss": 0.9329, "grad_norm": 0.2850522994995117, "learning_rate": 0.0002, "epoch": 0.20016406890894176, "step": 3050}, {"loss": 1.0777, "grad_norm": 0.36680465936660767, "learning_rate": 0.0002, "epoch": 0.2008203445447088, "step": 3060}, {"loss": 0.9042, "grad_norm": 0.2879831790924072, "learning_rate": 0.0002, "epoch": 0.2014766201804758, "step": 3070}, {"loss": 0.8925, "grad_norm": 0.29652512073516846, "learning_rate": 0.0002, "epoch": 0.20213289581624283, "step": 3080}, {"loss": 0.9928, "grad_norm": 0.3131853938102722, "learning_rate": 0.0002, "epoch": 0.20278917145200984, "step": 3090}, {"loss": 0.9968, "grad_norm": 0.34605351090431213, "learning_rate": 0.0002, "epoch": 0.20344544708777687, "step": 3100}, {"loss": 0.9701, "grad_norm": 0.30542251467704773, "learning_rate": 0.0002, "epoch": 0.20410172272354388, "step": 3110}, {"loss": 0.9483, "grad_norm": 0.28131186962127686, "learning_rate": 0.0002, "epoch": 0.2047579983593109, "step": 3120}, {"loss": 0.9332, "grad_norm": 0.257859468460083, "learning_rate": 0.0002, "epoch": 0.20541427399507795, "step": 3130}, {"loss": 0.9771, "grad_norm": 0.32994887232780457, "learning_rate": 0.0002, "epoch": 0.20607054963084495, "step": 3140}, {"loss": 0.9368, "grad_norm": 0.26293760538101196, "learning_rate": 0.0002, "epoch": 0.20672682526661199, "step": 3150}, {"loss": 0.9801, "grad_norm": 0.29514846205711365, "learning_rate": 0.0002, "epoch": 0.207383100902379, "step": 3160}, {"loss": 0.9879, "grad_norm": 0.36102691292762756, "learning_rate": 0.0002, "epoch": 0.20803937653814603, "step": 3170}, {"loss": 0.9322, "grad_norm": 0.3298998475074768, "learning_rate": 0.0002, "epoch": 0.20869565217391303, "step": 3180}, {"loss": 0.9456, "grad_norm": 0.29835769534111023, "learning_rate": 0.0002, "epoch": 0.20935192780968007, "step": 3190}, {"loss": 1.0199, "grad_norm": 0.3438013792037964, "learning_rate": 0.0002, "epoch": 0.2100082034454471, "step": 3200}, {"loss": 0.9104, "grad_norm": 0.2945845127105713, "learning_rate": 0.0002, "epoch": 0.2106644790812141, "step": 3210}, {"loss": 0.9231, "grad_norm": 0.3173643946647644, "learning_rate": 0.0002, "epoch": 0.21132075471698114, "step": 3220}, {"loss": 0.892, "grad_norm": 0.30580341815948486, "learning_rate": 0.0002, "epoch": 0.21197703035274815, "step": 3230}, {"loss": 0.9458, "grad_norm": 0.32913172245025635, "learning_rate": 0.0002, "epoch": 0.21263330598851518, "step": 3240}, {"loss": 0.9475, "grad_norm": 0.2739659249782562, "learning_rate": 0.0002, "epoch": 0.2132895816242822, "step": 3250}, {"loss": 0.9185, "grad_norm": 0.3016273081302643, "learning_rate": 0.0002, "epoch": 0.21394585726004922, "step": 3260}, {"loss": 0.9819, "grad_norm": 0.2818678319454193, "learning_rate": 0.0002, "epoch": 0.21460213289581626, "step": 3270}, {"loss": 0.9535, "grad_norm": 0.3243506848812103, "learning_rate": 0.0002, "epoch": 0.21525840853158326, "step": 3280}, {"loss": 0.9355, "grad_norm": 0.38120919466018677, "learning_rate": 0.0002, "epoch": 0.2159146841673503, "step": 3290}, {"loss": 0.9556, "grad_norm": 0.3041105568408966, "learning_rate": 0.0002, "epoch": 0.2165709598031173, "step": 3300}, {"loss": 0.9172, "grad_norm": 0.2648089528083801, "learning_rate": 0.0002, "epoch": 0.21722723543888434, "step": 3310}, {"loss": 0.9567, "grad_norm": 0.324095219373703, "learning_rate": 0.0002, "epoch": 0.21788351107465134, "step": 3320}, {"loss": 0.9586, "grad_norm": 0.2796897888183594, "learning_rate": 0.0002, "epoch": 0.21853978671041838, "step": 3330}, {"loss": 0.9505, "grad_norm": 0.30163177847862244, "learning_rate": 0.0002, "epoch": 0.2191960623461854, "step": 3340}, {"loss": 0.9515, "grad_norm": 0.29213520884513855, "learning_rate": 0.0002, "epoch": 0.21985233798195242, "step": 3350}, {"loss": 0.9692, "grad_norm": 0.28203412890434265, "learning_rate": 0.0002, "epoch": 0.22050861361771945, "step": 3360}, {"loss": 0.94, "grad_norm": 0.31636562943458557, "learning_rate": 0.0002, "epoch": 0.22116488925348646, "step": 3370}, {"loss": 0.901, "grad_norm": 0.3153952956199646, "learning_rate": 0.0002, "epoch": 0.2218211648892535, "step": 3380}, {"loss": 0.8994, "grad_norm": 0.2793780267238617, "learning_rate": 0.0002, "epoch": 0.2224774405250205, "step": 3390}, {"loss": 0.9424, "grad_norm": 0.36783504486083984, "learning_rate": 0.0002, "epoch": 0.22313371616078753, "step": 3400}, {"loss": 0.9837, "grad_norm": 0.31803956627845764, "learning_rate": 0.0002, "epoch": 0.22378999179655457, "step": 3410}, {"loss": 0.9234, "grad_norm": 0.31517738103866577, "learning_rate": 0.0002, "epoch": 0.22444626743232157, "step": 3420}, {"loss": 0.9336, "grad_norm": 0.4124458432197571, "learning_rate": 0.0002, "epoch": 0.2251025430680886, "step": 3430}, {"loss": 0.8587, "grad_norm": 0.29330259561538696, "learning_rate": 0.0002, "epoch": 0.2257588187038556, "step": 3440}, {"loss": 0.925, "grad_norm": 0.281255304813385, "learning_rate": 0.0002, "epoch": 0.22641509433962265, "step": 3450}, {"loss": 0.9797, "grad_norm": 0.3765242397785187, "learning_rate": 0.0002, "epoch": 0.22707136997538965, "step": 3460}, {"loss": 1.0201, "grad_norm": 0.33658838272094727, "learning_rate": 0.0002, "epoch": 0.2277276456111567, "step": 3470}, {"loss": 0.9255, "grad_norm": 0.2809208035469055, "learning_rate": 0.0002, "epoch": 0.22838392124692372, "step": 3480}, {"loss": 0.8948, "grad_norm": 0.3064846098423004, "learning_rate": 0.0002, "epoch": 0.22904019688269073, "step": 3490}, {"loss": 0.8805, "grad_norm": 0.31982484459877014, "learning_rate": 0.0002, "epoch": 0.22969647251845776, "step": 3500}, {"loss": 0.927, "grad_norm": 0.27903324365615845, "learning_rate": 0.0002, "epoch": 0.23035274815422477, "step": 3510}, {"loss": 0.9103, "grad_norm": 0.35411202907562256, "learning_rate": 0.0002, "epoch": 0.2310090237899918, "step": 3520}, {"loss": 0.9087, "grad_norm": 0.2796868681907654, "learning_rate": 0.0002, "epoch": 0.2316652994257588, "step": 3530}, {"loss": 0.9861, "grad_norm": 0.3428329825401306, "learning_rate": 0.0002, "epoch": 0.23232157506152584, "step": 3540}, {"loss": 0.9025, "grad_norm": 0.30563032627105713, "learning_rate": 0.0002, "epoch": 0.23297785069729288, "step": 3550}, {"loss": 0.9132, "grad_norm": 0.2954406142234802, "learning_rate": 0.0002, "epoch": 0.23363412633305988, "step": 3560}, {"loss": 0.9444, "grad_norm": 0.3328028917312622, "learning_rate": 0.0002, "epoch": 0.23429040196882692, "step": 3570}, {"loss": 0.9161, "grad_norm": 0.32020696997642517, "learning_rate": 0.0002, "epoch": 0.23494667760459392, "step": 3580}, {"loss": 0.9193, "grad_norm": 0.2774750292301178, "learning_rate": 0.0002, "epoch": 0.23560295324036096, "step": 3590}, {"loss": 0.8935, "grad_norm": 0.24560679495334625, "learning_rate": 0.0002, "epoch": 0.23625922887612796, "step": 3600}, {"loss": 0.9335, "grad_norm": 0.3278765082359314, "learning_rate": 0.0002, "epoch": 0.236915504511895, "step": 3610}, {"loss": 0.8958, "grad_norm": 0.4417719542980194, "learning_rate": 0.0002, "epoch": 0.23757178014766203, "step": 3620}, {"loss": 0.9683, "grad_norm": 0.3339618742465973, "learning_rate": 0.0002, "epoch": 0.23822805578342904, "step": 3630}, {"loss": 0.9536, "grad_norm": 0.2925402522087097, "learning_rate": 0.0002, "epoch": 0.23888433141919607, "step": 3640}, {"loss": 0.9695, "grad_norm": 0.3060242831707001, "learning_rate": 0.0002, "epoch": 0.23954060705496308, "step": 3650}, {"loss": 0.9573, "grad_norm": 0.3095077872276306, "learning_rate": 0.0002, "epoch": 0.2401968826907301, "step": 3660}, {"loss": 0.9023, "grad_norm": 0.31151828169822693, "learning_rate": 0.0002, "epoch": 0.24085315832649712, "step": 3670}, {"loss": 0.9642, "grad_norm": 0.293610543012619, "learning_rate": 0.0002, "epoch": 0.24150943396226415, "step": 3680}, {"loss": 0.9449, "grad_norm": 0.28868040442466736, "learning_rate": 0.0002, "epoch": 0.2421657095980312, "step": 3690}, {"loss": 0.9258, "grad_norm": 0.3123609721660614, "learning_rate": 0.0002, "epoch": 0.2428219852337982, "step": 3700}, {"loss": 0.9013, "grad_norm": 0.29155978560447693, "learning_rate": 0.0002, "epoch": 0.24347826086956523, "step": 3710}, {"loss": 0.8977, "grad_norm": 0.36799871921539307, "learning_rate": 0.0002, "epoch": 0.24413453650533223, "step": 3720}, {"loss": 1.0157, "grad_norm": 0.30915290117263794, "learning_rate": 0.0002, "epoch": 0.24479081214109927, "step": 3730}, {"loss": 0.9137, "grad_norm": 0.2889885902404785, "learning_rate": 0.0002, "epoch": 0.24544708777686627, "step": 3740}, {"loss": 0.9279, "grad_norm": 0.30213671922683716, "learning_rate": 0.0002, "epoch": 0.2461033634126333, "step": 3750}, {"loss": 0.9202, "grad_norm": 0.3242565393447876, "learning_rate": 0.0002, "epoch": 0.24675963904840031, "step": 3760}, {"loss": 0.9323, "grad_norm": 0.27826303243637085, "learning_rate": 0.0002, "epoch": 0.24741591468416735, "step": 3770}, {"loss": 0.933, "grad_norm": 0.2813768982887268, "learning_rate": 0.0002, "epoch": 0.24807219031993438, "step": 3780}, {"loss": 0.9046, "grad_norm": 0.34054139256477356, "learning_rate": 0.0002, "epoch": 0.2487284659557014, "step": 3790}, {"loss": 0.9252, "grad_norm": 0.2804257869720459, "learning_rate": 0.0002, "epoch": 0.24938474159146842, "step": 3800}, {"loss": 0.8305, "grad_norm": 0.2920171618461609, "learning_rate": 0.0002, "epoch": 0.25004101722723543, "step": 3810}, {"loss": 0.883, "grad_norm": 0.2967351973056793, "learning_rate": 0.0002, "epoch": 0.25069729286300246, "step": 3820}, {"loss": 0.8767, "grad_norm": 0.27417901158332825, "learning_rate": 0.0002, "epoch": 0.2513535684987695, "step": 3830}, {"loss": 0.9459, "grad_norm": 0.30916500091552734, "learning_rate": 0.0002, "epoch": 0.25200984413453653, "step": 3840}, {"loss": 0.9774, "grad_norm": 0.3281477391719818, "learning_rate": 0.0002, "epoch": 0.2526661197703035, "step": 3850}, {"loss": 0.8779, "grad_norm": 0.3417379856109619, "learning_rate": 0.0002, "epoch": 0.25332239540607054, "step": 3860}, {"loss": 0.9155, "grad_norm": 0.3119784891605377, "learning_rate": 0.0002, "epoch": 0.2539786710418376, "step": 3870}, {"loss": 0.9466, "grad_norm": 0.3053247332572937, "learning_rate": 0.0002, "epoch": 0.2546349466776046, "step": 3880}, {"loss": 0.8882, "grad_norm": 0.3155108094215393, "learning_rate": 0.0002, "epoch": 0.2552912223133716, "step": 3890}, {"loss": 0.9464, "grad_norm": 0.3240940272808075, "learning_rate": 0.0002, "epoch": 0.2559474979491386, "step": 3900}, {"loss": 0.9765, "grad_norm": 0.32433149218559265, "learning_rate": 0.0002, "epoch": 0.25660377358490566, "step": 3910}, {"loss": 0.9714, "grad_norm": 0.2982637286186218, "learning_rate": 0.0002, "epoch": 0.2572600492206727, "step": 3920}, {"loss": 0.8758, "grad_norm": 0.34742113947868347, "learning_rate": 0.0002, "epoch": 0.2579163248564397, "step": 3930}, {"loss": 0.9335, "grad_norm": 0.3415607213973999, "learning_rate": 0.0002, "epoch": 0.2585726004922067, "step": 3940}, {"loss": 0.9016, "grad_norm": 0.2852018475532532, "learning_rate": 0.0002, "epoch": 0.25922887612797374, "step": 3950}, {"loss": 0.9972, "grad_norm": 0.2924928665161133, "learning_rate": 0.0002, "epoch": 0.2598851517637408, "step": 3960}, {"loss": 0.9524, "grad_norm": 0.29387834668159485, "learning_rate": 0.0002, "epoch": 0.2605414273995078, "step": 3970}, {"loss": 0.9197, "grad_norm": 0.32400810718536377, "learning_rate": 0.0002, "epoch": 0.26119770303527484, "step": 3980}, {"loss": 0.9886, "grad_norm": 0.35629674792289734, "learning_rate": 0.0002, "epoch": 0.2618539786710418, "step": 3990}, {"loss": 0.9002, "grad_norm": 0.40292513370513916, "learning_rate": 0.0002, "epoch": 0.26251025430680885, "step": 4000}, {"loss": 0.9561, "grad_norm": 0.3452664613723755, "learning_rate": 0.0002, "epoch": 0.2631665299425759, "step": 4010}, {"loss": 0.9231, "grad_norm": 0.2956177592277527, "learning_rate": 0.0002, "epoch": 0.2638228055783429, "step": 4020}, {"loss": 0.9269, "grad_norm": 0.3159438669681549, "learning_rate": 0.0002, "epoch": 0.2644790812141099, "step": 4030}, {"loss": 0.9682, "grad_norm": 0.25804826617240906, "learning_rate": 0.0002, "epoch": 0.26513535684987694, "step": 4040}, {"loss": 0.9681, "grad_norm": 0.3446296751499176, "learning_rate": 0.0002, "epoch": 0.26579163248564397, "step": 4050}, {"loss": 0.9451, "grad_norm": 0.32568585872650146, "learning_rate": 0.0002, "epoch": 0.266447908121411, "step": 4060}, {"loss": 0.907, "grad_norm": 0.2872511148452759, "learning_rate": 0.0002, "epoch": 0.26710418375717804, "step": 4070}, {"loss": 0.9365, "grad_norm": 0.327495276927948, "learning_rate": 0.0002, "epoch": 0.267760459392945, "step": 4080}, {"loss": 0.9778, "grad_norm": 0.40853601694107056, "learning_rate": 0.0002, "epoch": 0.26841673502871205, "step": 4090}, {"loss": 0.9232, "grad_norm": 0.32097089290618896, "learning_rate": 0.0002, "epoch": 0.2690730106644791, "step": 4100}, {"loss": 0.9622, "grad_norm": 0.32956209778785706, "learning_rate": 0.0002, "epoch": 0.2697292863002461, "step": 4110}, {"loss": 0.911, "grad_norm": 0.30057013034820557, "learning_rate": 0.0002, "epoch": 0.27038556193601315, "step": 4120}, {"loss": 0.9045, "grad_norm": 0.30630001425743103, "learning_rate": 0.0002, "epoch": 0.27104183757178013, "step": 4130}, {"loss": 0.8914, "grad_norm": 0.31977957487106323, "learning_rate": 0.0002, "epoch": 0.27169811320754716, "step": 4140}, {"loss": 0.9619, "grad_norm": 0.29783955216407776, "learning_rate": 0.0002, "epoch": 0.2723543888433142, "step": 4150}, {"loss": 1.0299, "grad_norm": 0.37371826171875, "learning_rate": 0.0002, "epoch": 0.27301066447908123, "step": 4160}, {"loss": 0.9072, "grad_norm": 0.27656471729278564, "learning_rate": 0.0002, "epoch": 0.2736669401148482, "step": 4170}, {"loss": 0.873, "grad_norm": 0.29685157537460327, "learning_rate": 0.0002, "epoch": 0.27432321575061525, "step": 4180}, {"loss": 1.0057, "grad_norm": 0.3376981317996979, "learning_rate": 0.0002, "epoch": 0.2749794913863823, "step": 4190}, {"loss": 0.9671, "grad_norm": 0.2933194041252136, "learning_rate": 0.0002, "epoch": 0.2756357670221493, "step": 4200}, {"loss": 0.9359, "grad_norm": 0.27594974637031555, "learning_rate": 0.0002, "epoch": 0.27629204265791635, "step": 4210}, {"loss": 0.9336, "grad_norm": 0.29267510771751404, "learning_rate": 0.0002, "epoch": 0.2769483182936833, "step": 4220}, {"loss": 0.9241, "grad_norm": 0.3266076445579529, "learning_rate": 0.0002, "epoch": 0.27760459392945036, "step": 4230}, {"loss": 0.933, "grad_norm": 0.3626921474933624, "learning_rate": 0.0002, "epoch": 0.2782608695652174, "step": 4240}, {"loss": 0.9533, "grad_norm": 0.3043886423110962, "learning_rate": 0.0002, "epoch": 0.27891714520098443, "step": 4250}, {"loss": 0.902, "grad_norm": 0.28676939010620117, "learning_rate": 0.0002, "epoch": 0.27957342083675146, "step": 4260}, {"loss": 0.9565, "grad_norm": 0.32501107454299927, "learning_rate": 0.0002, "epoch": 0.28022969647251844, "step": 4270}, {"loss": 1.0009, "grad_norm": 0.3533550202846527, "learning_rate": 0.0002, "epoch": 0.2808859721082855, "step": 4280}, {"loss": 0.9128, "grad_norm": 0.28312650322914124, "learning_rate": 0.0002, "epoch": 0.2815422477440525, "step": 4290}, {"loss": 0.9626, "grad_norm": 0.2910906672477722, "learning_rate": 0.0002, "epoch": 0.28219852337981954, "step": 4300}, {"loss": 0.8803, "grad_norm": 0.27612248063087463, "learning_rate": 0.0002, "epoch": 0.2828547990155865, "step": 4310}, {"loss": 0.9451, "grad_norm": 0.416061133146286, "learning_rate": 0.0002, "epoch": 0.28351107465135356, "step": 4320}, {"loss": 0.9197, "grad_norm": 0.36900338530540466, "learning_rate": 0.0002, "epoch": 0.2841673502871206, "step": 4330}, {"loss": 0.9623, "grad_norm": 0.32224977016448975, "learning_rate": 0.0002, "epoch": 0.2848236259228876, "step": 4340}, {"loss": 0.9264, "grad_norm": 0.2805930972099304, "learning_rate": 0.0002, "epoch": 0.28547990155865466, "step": 4350}, {"loss": 0.9486, "grad_norm": 0.286539226770401, "learning_rate": 0.0002, "epoch": 0.28613617719442164, "step": 4360}, {"loss": 0.9225, "grad_norm": 0.3326348066329956, "learning_rate": 0.0002, "epoch": 0.28679245283018867, "step": 4370}, {"loss": 0.8774, "grad_norm": 0.29906603693962097, "learning_rate": 0.0002, "epoch": 0.2874487284659557, "step": 4380}, {"loss": 0.9582, "grad_norm": 0.33906206488609314, "learning_rate": 0.0002, "epoch": 0.28810500410172274, "step": 4390}, {"loss": 0.8987, "grad_norm": 0.36083030700683594, "learning_rate": 0.0002, "epoch": 0.2887612797374898, "step": 4400}, {"loss": 0.9189, "grad_norm": 0.30945461988449097, "learning_rate": 0.0002, "epoch": 0.28941755537325675, "step": 4410}, {"loss": 0.9454, "grad_norm": 0.29673129320144653, "learning_rate": 0.0002, "epoch": 0.2900738310090238, "step": 4420}, {"loss": 0.9454, "grad_norm": 0.29949837923049927, "learning_rate": 0.0002, "epoch": 0.2907301066447908, "step": 4430}, {"loss": 0.9026, "grad_norm": 0.29319390654563904, "learning_rate": 0.0002, "epoch": 0.29138638228055785, "step": 4440}, {"loss": 0.9318, "grad_norm": 0.38954150676727295, "learning_rate": 0.0002, "epoch": 0.29204265791632483, "step": 4450}, {"loss": 1.0558, "grad_norm": 0.32702240347862244, "learning_rate": 0.0002, "epoch": 0.29269893355209187, "step": 4460}, {"loss": 0.9684, "grad_norm": 0.3480510413646698, "learning_rate": 0.0002, "epoch": 0.2933552091878589, "step": 4470}, {"loss": 0.9267, "grad_norm": 0.34290337562561035, "learning_rate": 0.0002, "epoch": 0.29401148482362593, "step": 4480}, {"loss": 0.9561, "grad_norm": 0.3169049620628357, "learning_rate": 0.0002, "epoch": 0.29466776045939297, "step": 4490}, {"loss": 0.9386, "grad_norm": 0.3368853032588959, "learning_rate": 0.0002, "epoch": 0.29532403609515995, "step": 4500}, {"loss": 0.8803, "grad_norm": 0.30214768648147583, "learning_rate": 0.0002, "epoch": 0.295980311730927, "step": 4510}, {"loss": 0.9485, "grad_norm": 0.33051690459251404, "learning_rate": 0.0002, "epoch": 0.296636587366694, "step": 4520}, {"loss": 0.9498, "grad_norm": 0.37137898802757263, "learning_rate": 0.0002, "epoch": 0.29729286300246105, "step": 4530}, {"loss": 0.9518, "grad_norm": 0.3780321180820465, "learning_rate": 0.0002, "epoch": 0.29794913863822803, "step": 4540}, {"loss": 0.8912, "grad_norm": 0.3048851788043976, "learning_rate": 0.0002, "epoch": 0.29860541427399506, "step": 4550}, {"loss": 1.0054, "grad_norm": 0.36871910095214844, "learning_rate": 0.0002, "epoch": 0.2992616899097621, "step": 4560}, {"loss": 0.9172, "grad_norm": 0.3335227966308594, "learning_rate": 0.0002, "epoch": 0.29991796554552913, "step": 4570}, {"loss": 0.9368, "grad_norm": 0.3256683349609375, "learning_rate": 0.0002, "epoch": 0.30057424118129616, "step": 4580}, {"loss": 0.9641, "grad_norm": 0.34209194779396057, "learning_rate": 0.0002, "epoch": 0.30123051681706314, "step": 4590}, {"loss": 0.9128, "grad_norm": 0.278749018907547, "learning_rate": 0.0002, "epoch": 0.3018867924528302, "step": 4600}, {"loss": 0.931, "grad_norm": 0.28081533312797546, "learning_rate": 0.0002, "epoch": 0.3025430680885972, "step": 4610}, {"loss": 0.9084, "grad_norm": 0.35687389969825745, "learning_rate": 0.0002, "epoch": 0.30319934372436425, "step": 4620}, {"loss": 0.8927, "grad_norm": 0.367288738489151, "learning_rate": 0.0002, "epoch": 0.3038556193601313, "step": 4630}, {"loss": 0.9175, "grad_norm": 0.40433239936828613, "learning_rate": 0.0002, "epoch": 0.30451189499589826, "step": 4640}, {"loss": 0.9115, "grad_norm": 0.31489279866218567, "learning_rate": 0.0002, "epoch": 0.3051681706316653, "step": 4650}, {"loss": 0.9843, "grad_norm": 0.30823174118995667, "learning_rate": 0.0002, "epoch": 0.3058244462674323, "step": 4660}, {"loss": 0.964, "grad_norm": 0.3204546570777893, "learning_rate": 0.0002, "epoch": 0.30648072190319936, "step": 4670}, {"loss": 0.9754, "grad_norm": 0.3188243508338928, "learning_rate": 0.0002, "epoch": 0.30713699753896634, "step": 4680}, {"loss": 0.8923, "grad_norm": 0.31931981444358826, "learning_rate": 0.0002, "epoch": 0.3077932731747334, "step": 4690}, {"loss": 0.9113, "grad_norm": 0.33908605575561523, "learning_rate": 0.0002, "epoch": 0.3084495488105004, "step": 4700}, {"loss": 0.9552, "grad_norm": 0.3362937569618225, "learning_rate": 0.0002, "epoch": 0.30910582444626744, "step": 4710}, {"loss": 0.9695, "grad_norm": 0.3304995000362396, "learning_rate": 0.0002, "epoch": 0.3097621000820345, "step": 4720}, {"loss": 0.9588, "grad_norm": 0.39165404438972473, "learning_rate": 0.0002, "epoch": 0.31041837571780145, "step": 4730}, {"loss": 0.9583, "grad_norm": 0.29097774624824524, "learning_rate": 0.0002, "epoch": 0.3110746513535685, "step": 4740}, {"loss": 0.9224, "grad_norm": 0.289065420627594, "learning_rate": 0.0002, "epoch": 0.3117309269893355, "step": 4750}, {"loss": 0.8905, "grad_norm": 0.32188501954078674, "learning_rate": 0.0002, "epoch": 0.31238720262510256, "step": 4760}, {"loss": 0.951, "grad_norm": 0.2738335430622101, "learning_rate": 0.0002, "epoch": 0.3130434782608696, "step": 4770}, {"loss": 0.9662, "grad_norm": 0.2811134457588196, "learning_rate": 0.0002, "epoch": 0.31369975389663657, "step": 4780}, {"loss": 0.9372, "grad_norm": 0.2986338138580322, "learning_rate": 0.0002, "epoch": 0.3143560295324036, "step": 4790}, {"loss": 0.9016, "grad_norm": 0.30152231454849243, "learning_rate": 0.0002, "epoch": 0.31501230516817064, "step": 4800}, {"loss": 0.9902, "grad_norm": 0.37527933716773987, "learning_rate": 0.0002, "epoch": 0.31566858080393767, "step": 4810}, {"loss": 0.9786, "grad_norm": 0.4206887483596802, "learning_rate": 0.0002, "epoch": 0.31632485643970465, "step": 4820}, {"loss": 0.9265, "grad_norm": 0.3060953617095947, "learning_rate": 0.0002, "epoch": 0.3169811320754717, "step": 4830}, {"loss": 0.874, "grad_norm": 0.28399568796157837, "learning_rate": 0.0002, "epoch": 0.3176374077112387, "step": 4840}, {"loss": 0.9287, "grad_norm": 0.30498018860816956, "learning_rate": 0.0002, "epoch": 0.31829368334700575, "step": 4850}, {"loss": 0.9682, "grad_norm": 0.3168811798095703, "learning_rate": 0.0002, "epoch": 0.3189499589827728, "step": 4860}, {"loss": 0.9916, "grad_norm": 0.33051279187202454, "learning_rate": 0.0002, "epoch": 0.31960623461853976, "step": 4870}, {"loss": 0.9427, "grad_norm": 0.34563300013542175, "learning_rate": 0.0002, "epoch": 0.3202625102543068, "step": 4880}, {"loss": 0.9018, "grad_norm": 0.28437602519989014, "learning_rate": 0.0002, "epoch": 0.32091878589007383, "step": 4890}, {"loss": 0.9357, "grad_norm": 0.3883301913738251, "learning_rate": 0.0002, "epoch": 0.32157506152584087, "step": 4900}, {"loss": 0.9034, "grad_norm": 0.2933879792690277, "learning_rate": 0.0002, "epoch": 0.3222313371616079, "step": 4910}, {"loss": 0.9332, "grad_norm": 0.3490493595600128, "learning_rate": 0.0002, "epoch": 0.3228876127973749, "step": 4920}, {"loss": 0.9424, "grad_norm": 0.29370415210723877, "learning_rate": 0.0002, "epoch": 0.3235438884331419, "step": 4930}, {"loss": 0.9751, "grad_norm": 0.32653453946113586, "learning_rate": 0.0002, "epoch": 0.32420016406890895, "step": 4940}, {"loss": 0.9278, "grad_norm": 0.32540783286094666, "learning_rate": 0.0002, "epoch": 0.324856439704676, "step": 4950}, {"loss": 0.9309, "grad_norm": 0.37576451897621155, "learning_rate": 0.0002, "epoch": 0.32551271534044296, "step": 4960}, {"loss": 0.9612, "grad_norm": 0.27892962098121643, "learning_rate": 0.0002, "epoch": 0.32616899097621, "step": 4970}, {"loss": 0.9765, "grad_norm": 0.38215309381484985, "learning_rate": 0.0002, "epoch": 0.32682526661197703, "step": 4980}, {"loss": 0.9226, "grad_norm": 0.4062992334365845, "learning_rate": 0.0002, "epoch": 0.32748154224774406, "step": 4990}, {"loss": 0.9157, "grad_norm": 0.3053376376628876, "learning_rate": 0.0002, "epoch": 0.3281378178835111, "step": 5000}, {"loss": 0.9117, "grad_norm": 0.3228836953639984, "learning_rate": 0.0002, "epoch": 0.3287940935192781, "step": 5010}, {"loss": 0.8644, "grad_norm": 0.27714234590530396, "learning_rate": 0.0002, "epoch": 0.3294503691550451, "step": 5020}, {"loss": 0.9034, "grad_norm": 0.3127416968345642, "learning_rate": 0.0002, "epoch": 0.33010664479081214, "step": 5030}, {"loss": 0.89, "grad_norm": 0.3653234839439392, "learning_rate": 0.0002, "epoch": 0.3307629204265792, "step": 5040}, {"loss": 0.9313, "grad_norm": 0.33517029881477356, "learning_rate": 0.0002, "epoch": 0.3314191960623462, "step": 5050}, {"loss": 0.9219, "grad_norm": 0.3412803113460541, "learning_rate": 0.0002, "epoch": 0.3320754716981132, "step": 5060}, {"loss": 0.9358, "grad_norm": 0.32958096265792847, "learning_rate": 0.0002, "epoch": 0.3327317473338802, "step": 5070}, {"loss": 0.9479, "grad_norm": 0.2991558015346527, "learning_rate": 0.0002, "epoch": 0.33338802296964726, "step": 5080}, {"loss": 0.9754, "grad_norm": 0.35870906710624695, "learning_rate": 0.0002, "epoch": 0.3340442986054143, "step": 5090}, {"loss": 0.9183, "grad_norm": 0.26751458644866943, "learning_rate": 0.0002, "epoch": 0.33470057424118127, "step": 5100}, {"loss": 0.9676, "grad_norm": 0.35588568449020386, "learning_rate": 0.0002, "epoch": 0.3353568498769483, "step": 5110}, {"loss": 0.8964, "grad_norm": 0.41509315371513367, "learning_rate": 0.0002, "epoch": 0.33601312551271534, "step": 5120}, {"loss": 0.9864, "grad_norm": 0.37643107771873474, "learning_rate": 0.0002, "epoch": 0.3366694011484824, "step": 5130}, {"loss": 0.9367, "grad_norm": 0.3062657117843628, "learning_rate": 0.0002, "epoch": 0.3373256767842494, "step": 5140}, {"loss": 0.9167, "grad_norm": 0.2778759300708771, "learning_rate": 0.0002, "epoch": 0.3379819524200164, "step": 5150}, {"loss": 0.9787, "grad_norm": 0.40632501244544983, "learning_rate": 0.0002, "epoch": 0.3386382280557834, "step": 5160}, {"loss": 0.8941, "grad_norm": 0.2996899485588074, "learning_rate": 0.0002, "epoch": 0.33929450369155045, "step": 5170}, {"loss": 0.9611, "grad_norm": 0.2869918942451477, "learning_rate": 0.0002, "epoch": 0.3399507793273175, "step": 5180}, {"loss": 0.9201, "grad_norm": 0.3303343951702118, "learning_rate": 0.0002, "epoch": 0.3406070549630845, "step": 5190}, {"loss": 0.9162, "grad_norm": 0.3582284450531006, "learning_rate": 0.0002, "epoch": 0.3412633305988515, "step": 5200}, {"loss": 0.9441, "grad_norm": 0.3600800037384033, "learning_rate": 0.0002, "epoch": 0.34191960623461853, "step": 5210}, {"loss": 0.9061, "grad_norm": 0.3147357106208801, "learning_rate": 0.0002, "epoch": 0.34257588187038557, "step": 5220}, {"loss": 0.9189, "grad_norm": 0.3327115774154663, "learning_rate": 0.0002, "epoch": 0.3432321575061526, "step": 5230}, {"loss": 0.9606, "grad_norm": 0.3514555096626282, "learning_rate": 0.0002, "epoch": 0.3438884331419196, "step": 5240}, {"loss": 0.8944, "grad_norm": 0.35748785734176636, "learning_rate": 0.0002, "epoch": 0.3445447087776866, "step": 5250}, {"loss": 0.9228, "grad_norm": 0.3179738223552704, "learning_rate": 0.0002, "epoch": 0.34520098441345365, "step": 5260}, {"loss": 0.9313, "grad_norm": 0.3399673104286194, "learning_rate": 0.0002, "epoch": 0.3458572600492207, "step": 5270}, {"loss": 0.9379, "grad_norm": 0.32293835282325745, "learning_rate": 0.0002, "epoch": 0.3465135356849877, "step": 5280}, {"loss": 0.9066, "grad_norm": 0.3028234839439392, "learning_rate": 0.0002, "epoch": 0.3471698113207547, "step": 5290}, {"loss": 0.929, "grad_norm": 0.3056369423866272, "learning_rate": 0.0002, "epoch": 0.34782608695652173, "step": 5300}, {"loss": 0.9597, "grad_norm": 0.36918318271636963, "learning_rate": 0.0002, "epoch": 0.34848236259228876, "step": 5310}, {"loss": 0.9365, "grad_norm": 0.2661174535751343, "learning_rate": 0.0002, "epoch": 0.3491386382280558, "step": 5320}, {"loss": 0.8832, "grad_norm": 0.294240266084671, "learning_rate": 0.0002, "epoch": 0.34979491386382283, "step": 5330}, {"loss": 0.9135, "grad_norm": 0.3509284555912018, "learning_rate": 0.0002, "epoch": 0.3504511894995898, "step": 5340}, {"loss": 0.8919, "grad_norm": 0.331851989030838, "learning_rate": 0.0002, "epoch": 0.35110746513535684, "step": 5350}, {"loss": 0.9357, "grad_norm": 0.39378881454467773, "learning_rate": 0.0002, "epoch": 0.3517637407711239, "step": 5360}, {"loss": 0.8844, "grad_norm": 0.3014044761657715, "learning_rate": 0.0002, "epoch": 0.3524200164068909, "step": 5370}, {"loss": 0.9234, "grad_norm": 0.3350718021392822, "learning_rate": 0.0002, "epoch": 0.3530762920426579, "step": 5380}, {"loss": 0.9119, "grad_norm": 0.29478874802589417, "learning_rate": 0.0002, "epoch": 0.3537325676784249, "step": 5390}, {"loss": 0.8781, "grad_norm": 0.3219947814941406, "learning_rate": 0.0002, "epoch": 0.35438884331419196, "step": 5400}, {"loss": 0.9527, "grad_norm": 0.31258803606033325, "learning_rate": 0.0002, "epoch": 0.355045118949959, "step": 5410}, {"loss": 0.9124, "grad_norm": 0.2951710820198059, "learning_rate": 0.0002, "epoch": 0.355701394585726, "step": 5420}, {"loss": 0.8997, "grad_norm": 0.3726331293582916, "learning_rate": 0.0002, "epoch": 0.356357670221493, "step": 5430}, {"loss": 0.9288, "grad_norm": 0.3571377098560333, "learning_rate": 0.0002, "epoch": 0.35701394585726004, "step": 5440}, {"loss": 0.9538, "grad_norm": 0.3004431128501892, "learning_rate": 0.0002, "epoch": 0.3576702214930271, "step": 5450}, {"loss": 0.9466, "grad_norm": 0.3258959949016571, "learning_rate": 0.0002, "epoch": 0.3583264971287941, "step": 5460}, {"loss": 0.8999, "grad_norm": 0.31265372037887573, "learning_rate": 0.0002, "epoch": 0.35898277276456114, "step": 5470}, {"loss": 0.9438, "grad_norm": 0.3120972514152527, "learning_rate": 0.0002, "epoch": 0.3596390484003281, "step": 5480}, {"loss": 0.9541, "grad_norm": 0.31449300050735474, "learning_rate": 0.0002, "epoch": 0.36029532403609515, "step": 5490}, {"loss": 0.9569, "grad_norm": 0.378205806016922, "learning_rate": 0.0002, "epoch": 0.3609515996718622, "step": 5500}, {"loss": 0.8964, "grad_norm": 0.34647420048713684, "learning_rate": 0.0002, "epoch": 0.3616078753076292, "step": 5510}, {"loss": 0.9965, "grad_norm": 0.34195807576179504, "learning_rate": 0.0002, "epoch": 0.3622641509433962, "step": 5520}, {"loss": 0.8916, "grad_norm": 0.35365694761276245, "learning_rate": 0.0002, "epoch": 0.36292042657916324, "step": 5530}, {"loss": 0.9429, "grad_norm": 0.3188548982143402, "learning_rate": 0.0002, "epoch": 0.36357670221493027, "step": 5540}, {"loss": 0.9258, "grad_norm": 0.35012900829315186, "learning_rate": 0.0002, "epoch": 0.3642329778506973, "step": 5550}, {"loss": 0.9338, "grad_norm": 0.3564446270465851, "learning_rate": 0.0002, "epoch": 0.36488925348646434, "step": 5560}, {"loss": 0.9531, "grad_norm": 0.29641905426979065, "learning_rate": 0.0002, "epoch": 0.3655455291222313, "step": 5570}, {"loss": 0.9507, "grad_norm": 0.3476453721523285, "learning_rate": 0.0002, "epoch": 0.36620180475799835, "step": 5580}, {"loss": 0.9509, "grad_norm": 0.32947802543640137, "learning_rate": 0.0002, "epoch": 0.3668580803937654, "step": 5590}, {"loss": 0.9577, "grad_norm": 0.28639012575149536, "learning_rate": 0.0002, "epoch": 0.3675143560295324, "step": 5600}, {"loss": 0.9196, "grad_norm": 0.35645192861557007, "learning_rate": 0.0002, "epoch": 0.36817063166529945, "step": 5610}, {"loss": 0.9544, "grad_norm": 0.3594934046268463, "learning_rate": 0.0002, "epoch": 0.36882690730106643, "step": 5620}, {"loss": 0.9081, "grad_norm": 0.3584282398223877, "learning_rate": 0.0002, "epoch": 0.36948318293683347, "step": 5630}, {"loss": 0.955, "grad_norm": 0.32094448804855347, "learning_rate": 0.0002, "epoch": 0.3701394585726005, "step": 5640}, {"loss": 0.9271, "grad_norm": 0.3229917287826538, "learning_rate": 0.0002, "epoch": 0.37079573420836753, "step": 5650}, {"loss": 0.933, "grad_norm": 0.3371448218822479, "learning_rate": 0.0002, "epoch": 0.3714520098441345, "step": 5660}, {"loss": 0.9189, "grad_norm": 0.38639435172080994, "learning_rate": 0.0002, "epoch": 0.37210828547990155, "step": 5670}, {"loss": 0.8987, "grad_norm": 0.37710604071617126, "learning_rate": 0.0002, "epoch": 0.3727645611156686, "step": 5680}, {"loss": 0.9133, "grad_norm": 0.28154921531677246, "learning_rate": 0.0002, "epoch": 0.3734208367514356, "step": 5690}, {"loss": 0.9104, "grad_norm": 0.35519927740097046, "learning_rate": 0.0002, "epoch": 0.37407711238720265, "step": 5700}, {"loss": 0.9187, "grad_norm": 0.3271195888519287, "learning_rate": 0.0002, "epoch": 0.3747333880229696, "step": 5710}, {"loss": 1.0015, "grad_norm": 0.30836206674575806, "learning_rate": 0.0002, "epoch": 0.37538966365873666, "step": 5720}, {"loss": 0.8888, "grad_norm": 0.36397358775138855, "learning_rate": 0.0002, "epoch": 0.3760459392945037, "step": 5730}, {"loss": 0.8962, "grad_norm": 0.31649520993232727, "learning_rate": 0.0002, "epoch": 0.37670221493027073, "step": 5740}, {"loss": 0.9667, "grad_norm": 0.3790256977081299, "learning_rate": 0.0002, "epoch": 0.37735849056603776, "step": 5750}, {"loss": 0.9548, "grad_norm": 0.34175634384155273, "learning_rate": 0.0002, "epoch": 0.37801476620180474, "step": 5760}, {"loss": 0.9031, "grad_norm": 0.4169650673866272, "learning_rate": 0.0002, "epoch": 0.3786710418375718, "step": 5770}, {"loss": 0.9279, "grad_norm": 0.3118833899497986, "learning_rate": 0.0002, "epoch": 0.3793273174733388, "step": 5780}, {"loss": 0.9204, "grad_norm": 0.3474520444869995, "learning_rate": 0.0002, "epoch": 0.37998359310910584, "step": 5790}, {"loss": 0.9845, "grad_norm": 0.40819892287254333, "learning_rate": 0.0002, "epoch": 0.3806398687448728, "step": 5800}, {"loss": 0.8863, "grad_norm": 0.34402596950531006, "learning_rate": 0.0002, "epoch": 0.38129614438063986, "step": 5810}, {"loss": 0.9356, "grad_norm": 0.31899532675743103, "learning_rate": 0.0002, "epoch": 0.3819524200164069, "step": 5820}, {"loss": 0.857, "grad_norm": 0.38860151171684265, "learning_rate": 0.0002, "epoch": 0.3826086956521739, "step": 5830}, {"loss": 0.9433, "grad_norm": 0.2960244417190552, "learning_rate": 0.0002, "epoch": 0.38326497128794096, "step": 5840}, {"loss": 0.9546, "grad_norm": 0.3614438772201538, "learning_rate": 0.0002, "epoch": 0.38392124692370794, "step": 5850}, {"loss": 0.9447, "grad_norm": 0.33000093698501587, "learning_rate": 0.0002, "epoch": 0.38457752255947497, "step": 5860}, {"loss": 0.964, "grad_norm": 0.3543432354927063, "learning_rate": 0.0002, "epoch": 0.385233798195242, "step": 5870}, {"loss": 0.9329, "grad_norm": 0.3281349837779999, "learning_rate": 0.0002, "epoch": 0.38589007383100904, "step": 5880}, {"loss": 0.9463, "grad_norm": 0.3775436580181122, "learning_rate": 0.0002, "epoch": 0.3865463494667761, "step": 5890}, {"loss": 0.9035, "grad_norm": 0.3037714660167694, "learning_rate": 0.0002, "epoch": 0.38720262510254305, "step": 5900}, {"loss": 0.8953, "grad_norm": 0.30885955691337585, "learning_rate": 0.0002, "epoch": 0.3878589007383101, "step": 5910}, {"loss": 0.8789, "grad_norm": 0.2839881181716919, "learning_rate": 0.0002, "epoch": 0.3885151763740771, "step": 5920}, {"loss": 0.9585, "grad_norm": 0.30544501543045044, "learning_rate": 0.0002, "epoch": 0.38917145200984415, "step": 5930}, {"loss": 0.9163, "grad_norm": 0.281946063041687, "learning_rate": 0.0002, "epoch": 0.38982772764561113, "step": 5940}, {"loss": 0.8999, "grad_norm": 0.44500306248664856, "learning_rate": 0.0002, "epoch": 0.39048400328137817, "step": 5950}, {"loss": 0.8928, "grad_norm": 0.324260950088501, "learning_rate": 0.0002, "epoch": 0.3911402789171452, "step": 5960}, {"loss": 0.8844, "grad_norm": 0.31535372138023376, "learning_rate": 0.0002, "epoch": 0.39179655455291224, "step": 5970}, {"loss": 0.9185, "grad_norm": 0.31681323051452637, "learning_rate": 0.0002, "epoch": 0.39245283018867927, "step": 5980}, {"loss": 0.9258, "grad_norm": 0.32668736577033997, "learning_rate": 0.0002, "epoch": 0.39310910582444625, "step": 5990}, {"loss": 0.9278, "grad_norm": 0.34531155228614807, "learning_rate": 0.0002, "epoch": 0.3937653814602133, "step": 6000}, {"loss": 0.8969, "grad_norm": 0.32312485575675964, "learning_rate": 0.0002, "epoch": 0.3944216570959803, "step": 6010}, {"loss": 0.9501, "grad_norm": 0.31243696808815, "learning_rate": 0.0002, "epoch": 0.39507793273174735, "step": 6020}, {"loss": 0.9174, "grad_norm": 0.3558828830718994, "learning_rate": 0.0002, "epoch": 0.39573420836751433, "step": 6030}, {"loss": 0.9486, "grad_norm": 0.3894706964492798, "learning_rate": 0.0002, "epoch": 0.39639048400328136, "step": 6040}, {"loss": 0.9549, "grad_norm": 0.36445918679237366, "learning_rate": 0.0002, "epoch": 0.3970467596390484, "step": 6050}, {"loss": 0.9268, "grad_norm": 0.34586355090141296, "learning_rate": 0.0002, "epoch": 0.39770303527481543, "step": 6060}, {"loss": 0.9256, "grad_norm": 0.4428504705429077, "learning_rate": 0.0002, "epoch": 0.39835931091058246, "step": 6070}, {"loss": 0.9706, "grad_norm": 0.32784751057624817, "learning_rate": 0.0002, "epoch": 0.39901558654634944, "step": 6080}, {"loss": 0.9107, "grad_norm": 0.30480754375457764, "learning_rate": 0.0002, "epoch": 0.3996718621821165, "step": 6090}, {"loss": 0.9431, "grad_norm": 0.3382161557674408, "learning_rate": 0.0002, "epoch": 0.4003281378178835, "step": 6100}, {"loss": 0.8669, "grad_norm": 0.34486305713653564, "learning_rate": 0.0002, "epoch": 0.40098441345365055, "step": 6110}, {"loss": 0.9174, "grad_norm": 0.3488329350948334, "learning_rate": 0.0002, "epoch": 0.4016406890894176, "step": 6120}, {"loss": 0.8907, "grad_norm": 0.2977526783943176, "learning_rate": 0.0002, "epoch": 0.40229696472518456, "step": 6130}, {"loss": 0.9088, "grad_norm": 0.27584630250930786, "learning_rate": 0.0002, "epoch": 0.4029532403609516, "step": 6140}, {"loss": 0.9062, "grad_norm": 0.30325573682785034, "learning_rate": 0.0002, "epoch": 0.4036095159967186, "step": 6150}, {"loss": 0.9626, "grad_norm": 0.3318769633769989, "learning_rate": 0.0002, "epoch": 0.40426579163248566, "step": 6160}, {"loss": 0.8849, "grad_norm": 0.27929258346557617, "learning_rate": 0.0002, "epoch": 0.40492206726825264, "step": 6170}, {"loss": 0.9118, "grad_norm": 0.3127989172935486, "learning_rate": 0.0002, "epoch": 0.4055783429040197, "step": 6180}, {"loss": 0.9659, "grad_norm": 0.34189802408218384, "learning_rate": 0.0002, "epoch": 0.4062346185397867, "step": 6190}, {"loss": 0.9041, "grad_norm": 0.2901368737220764, "learning_rate": 0.0002, "epoch": 0.40689089417555374, "step": 6200}, {"loss": 0.8863, "grad_norm": 0.2957440912723541, "learning_rate": 0.0002, "epoch": 0.4075471698113208, "step": 6210}, {"loss": 0.8872, "grad_norm": 0.35854384303092957, "learning_rate": 0.0002, "epoch": 0.40820344544708775, "step": 6220}, {"loss": 0.9913, "grad_norm": 0.3516639471054077, "learning_rate": 0.0002, "epoch": 0.4088597210828548, "step": 6230}, {"loss": 0.9526, "grad_norm": 0.33773764967918396, "learning_rate": 0.0002, "epoch": 0.4095159967186218, "step": 6240}, {"loss": 0.9116, "grad_norm": 0.3359714448451996, "learning_rate": 0.0002, "epoch": 0.41017227235438886, "step": 6250}, {"loss": 0.8789, "grad_norm": 0.37056079506874084, "learning_rate": 0.0002, "epoch": 0.4108285479901559, "step": 6260}, {"loss": 0.9262, "grad_norm": 0.37948423624038696, "learning_rate": 0.0002, "epoch": 0.41148482362592287, "step": 6270}, {"loss": 0.9333, "grad_norm": 0.30971240997314453, "learning_rate": 0.0002, "epoch": 0.4121410992616899, "step": 6280}, {"loss": 0.8636, "grad_norm": 0.3357695937156677, "learning_rate": 0.0002, "epoch": 0.41279737489745694, "step": 6290}, {"loss": 0.8857, "grad_norm": 0.2873750627040863, "learning_rate": 0.0002, "epoch": 0.41345365053322397, "step": 6300}, {"loss": 0.9401, "grad_norm": 0.31098586320877075, "learning_rate": 0.0002, "epoch": 0.41410992616899095, "step": 6310}, {"loss": 0.9113, "grad_norm": 0.3232232630252838, "learning_rate": 0.0002, "epoch": 0.414766201804758, "step": 6320}, {"loss": 0.9165, "grad_norm": 0.3094746172428131, "learning_rate": 0.0002, "epoch": 0.415422477440525, "step": 6330}, {"loss": 0.92, "grad_norm": 0.3032565414905548, "learning_rate": 0.0002, "epoch": 0.41607875307629205, "step": 6340}, {"loss": 0.9254, "grad_norm": 0.30271366238594055, "learning_rate": 0.0002, "epoch": 0.4167350287120591, "step": 6350}, {"loss": 0.9766, "grad_norm": 0.35404741764068604, "learning_rate": 0.0002, "epoch": 0.41739130434782606, "step": 6360}, {"loss": 0.889, "grad_norm": 0.33428773283958435, "learning_rate": 0.0002, "epoch": 0.4180475799835931, "step": 6370}, {"loss": 0.9284, "grad_norm": 0.3523367941379547, "learning_rate": 0.0002, "epoch": 0.41870385561936013, "step": 6380}, {"loss": 0.8831, "grad_norm": 0.28411954641342163, "learning_rate": 0.0002, "epoch": 0.41936013125512717, "step": 6390}, {"loss": 0.9429, "grad_norm": 0.3410111963748932, "learning_rate": 0.0002, "epoch": 0.4200164068908942, "step": 6400}, {"loss": 0.8956, "grad_norm": 0.33837398886680603, "learning_rate": 0.0002, "epoch": 0.4206726825266612, "step": 6410}, {"loss": 0.9169, "grad_norm": 0.38739341497421265, "learning_rate": 0.0002, "epoch": 0.4213289581624282, "step": 6420}, {"loss": 0.8943, "grad_norm": 0.28810667991638184, "learning_rate": 0.0002, "epoch": 0.42198523379819525, "step": 6430}, {"loss": 0.9505, "grad_norm": 0.3242197036743164, "learning_rate": 0.0002, "epoch": 0.4226415094339623, "step": 6440}, {"loss": 0.9409, "grad_norm": 0.39538851380348206, "learning_rate": 0.0002, "epoch": 0.42329778506972926, "step": 6450}, {"loss": 0.9152, "grad_norm": 0.36792051792144775, "learning_rate": 0.0002, "epoch": 0.4239540607054963, "step": 6460}, {"loss": 0.9203, "grad_norm": 0.31599462032318115, "learning_rate": 0.0002, "epoch": 0.42461033634126333, "step": 6470}, {"loss": 0.926, "grad_norm": 0.3358552157878876, "learning_rate": 0.0002, "epoch": 0.42526661197703036, "step": 6480}, {"loss": 0.9337, "grad_norm": 0.40409335494041443, "learning_rate": 0.0002, "epoch": 0.4259228876127974, "step": 6490}, {"loss": 0.901, "grad_norm": 0.3785623610019684, "learning_rate": 0.0002, "epoch": 0.4265791632485644, "step": 6500}, {"loss": 0.9281, "grad_norm": 0.43033692240715027, "learning_rate": 0.0002, "epoch": 0.4272354388843314, "step": 6510}, {"loss": 0.9332, "grad_norm": 0.3074797987937927, "learning_rate": 0.0002, "epoch": 0.42789171452009844, "step": 6520}, {"loss": 0.9161, "grad_norm": 0.3498784005641937, "learning_rate": 0.0002, "epoch": 0.4285479901558655, "step": 6530}, {"loss": 0.9829, "grad_norm": 0.40507981181144714, "learning_rate": 0.0002, "epoch": 0.4292042657916325, "step": 6540}, {"loss": 0.9269, "grad_norm": 0.35614013671875, "learning_rate": 0.0002, "epoch": 0.4298605414273995, "step": 6550}, {"loss": 0.9033, "grad_norm": 0.451616495847702, "learning_rate": 0.0002, "epoch": 0.4305168170631665, "step": 6560}, {"loss": 0.9723, "grad_norm": 0.3707262873649597, "learning_rate": 0.0002, "epoch": 0.43117309269893356, "step": 6570}, {"loss": 0.9512, "grad_norm": 0.3271283209323883, "learning_rate": 0.0002, "epoch": 0.4318293683347006, "step": 6580}, {"loss": 0.8529, "grad_norm": 0.38105711340904236, "learning_rate": 0.0002, "epoch": 0.43248564397046757, "step": 6590}, {"loss": 0.9208, "grad_norm": 0.345217227935791, "learning_rate": 0.0002, "epoch": 0.4331419196062346, "step": 6600}, {"loss": 0.8958, "grad_norm": 0.3507174551486969, "learning_rate": 0.0002, "epoch": 0.43379819524200164, "step": 6610}, {"loss": 0.9402, "grad_norm": 0.318654328584671, "learning_rate": 0.0002, "epoch": 0.4344544708777687, "step": 6620}, {"loss": 0.9068, "grad_norm": 0.3628910183906555, "learning_rate": 0.0002, "epoch": 0.4351107465135357, "step": 6630}, {"loss": 0.9122, "grad_norm": 0.4454437494277954, "learning_rate": 0.0002, "epoch": 0.4357670221493027, "step": 6640}, {"loss": 1.0092, "grad_norm": 0.36227893829345703, "learning_rate": 0.0002, "epoch": 0.4364232977850697, "step": 6650}, {"loss": 0.9289, "grad_norm": 0.33995604515075684, "learning_rate": 0.0002, "epoch": 0.43707957342083675, "step": 6660}, {"loss": 0.9032, "grad_norm": 0.3784409165382385, "learning_rate": 0.0002, "epoch": 0.4377358490566038, "step": 6670}, {"loss": 0.9399, "grad_norm": 0.39470377564430237, "learning_rate": 0.0002, "epoch": 0.4383921246923708, "step": 6680}, {"loss": 0.8795, "grad_norm": 0.34198591113090515, "learning_rate": 0.0002, "epoch": 0.4390484003281378, "step": 6690}, {"loss": 0.9665, "grad_norm": 0.35575005412101746, "learning_rate": 0.0002, "epoch": 0.43970467596390483, "step": 6700}, {"loss": 0.9133, "grad_norm": 0.40915772318840027, "learning_rate": 0.0002, "epoch": 0.44036095159967187, "step": 6710}, {"loss": 0.9112, "grad_norm": 0.3526946008205414, "learning_rate": 0.0002, "epoch": 0.4410172272354389, "step": 6720}, {"loss": 0.8886, "grad_norm": 0.30411285161972046, "learning_rate": 0.0002, "epoch": 0.4416735028712059, "step": 6730}, {"loss": 0.8969, "grad_norm": 0.2935996949672699, "learning_rate": 0.0002, "epoch": 0.4423297785069729, "step": 6740}, {"loss": 0.8864, "grad_norm": 0.3806987702846527, "learning_rate": 0.0002, "epoch": 0.44298605414273995, "step": 6750}, {"loss": 0.9267, "grad_norm": 0.3329904079437256, "learning_rate": 0.0002, "epoch": 0.443642329778507, "step": 6760}, {"loss": 0.9223, "grad_norm": 0.33832767605781555, "learning_rate": 0.0002, "epoch": 0.444298605414274, "step": 6770}, {"loss": 0.9127, "grad_norm": 0.3387627899646759, "learning_rate": 0.0002, "epoch": 0.444954881050041, "step": 6780}, {"loss": 0.8853, "grad_norm": 0.32709822058677673, "learning_rate": 0.0002, "epoch": 0.44561115668580803, "step": 6790}, {"loss": 0.9059, "grad_norm": 0.2914820909500122, "learning_rate": 0.0002, "epoch": 0.44626743232157506, "step": 6800}, {"loss": 0.9209, "grad_norm": 0.3288695216178894, "learning_rate": 0.0002, "epoch": 0.4469237079573421, "step": 6810}, {"loss": 0.9208, "grad_norm": 0.5003459453582764, "learning_rate": 0.0002, "epoch": 0.44757998359310913, "step": 6820}, {"loss": 0.9797, "grad_norm": 0.359402060508728, "learning_rate": 0.0002, "epoch": 0.4482362592288761, "step": 6830}, {"loss": 0.856, "grad_norm": 0.35508984327316284, "learning_rate": 0.0002, "epoch": 0.44889253486464314, "step": 6840}, {"loss": 0.9513, "grad_norm": 0.33327680826187134, "learning_rate": 0.0002, "epoch": 0.4495488105004102, "step": 6850}, {"loss": 0.965, "grad_norm": 0.4084452688694, "learning_rate": 0.0002, "epoch": 0.4502050861361772, "step": 6860}, {"loss": 0.899, "grad_norm": 0.30232545733451843, "learning_rate": 0.0002, "epoch": 0.4508613617719442, "step": 6870}, {"loss": 0.9294, "grad_norm": 0.3244702219963074, "learning_rate": 0.0002, "epoch": 0.4515176374077112, "step": 6880}, {"loss": 0.9312, "grad_norm": 0.32148104906082153, "learning_rate": 0.0002, "epoch": 0.45217391304347826, "step": 6890}, {"loss": 1.007, "grad_norm": 0.3831937313079834, "learning_rate": 0.0002, "epoch": 0.4528301886792453, "step": 6900}, {"loss": 0.9111, "grad_norm": 0.33381497859954834, "learning_rate": 0.0002, "epoch": 0.4534864643150123, "step": 6910}, {"loss": 0.8955, "grad_norm": 0.4074220359325409, "learning_rate": 0.0002, "epoch": 0.4541427399507793, "step": 6920}, {"loss": 0.9326, "grad_norm": 0.35602903366088867, "learning_rate": 0.0002, "epoch": 0.45479901558654634, "step": 6930}, {"loss": 0.9146, "grad_norm": 0.3411916196346283, "learning_rate": 0.0002, "epoch": 0.4554552912223134, "step": 6940}, {"loss": 0.9382, "grad_norm": 0.38504868745803833, "learning_rate": 0.0002, "epoch": 0.4561115668580804, "step": 6950}, {"loss": 0.9105, "grad_norm": 0.5072926878929138, "learning_rate": 0.0002, "epoch": 0.45676784249384744, "step": 6960}, {"loss": 0.9383, "grad_norm": 0.38257330656051636, "learning_rate": 0.0002, "epoch": 0.4574241181296144, "step": 6970}, {"loss": 0.8864, "grad_norm": 0.40472176671028137, "learning_rate": 0.0002, "epoch": 0.45808039376538146, "step": 6980}, {"loss": 0.949, "grad_norm": 0.34501466155052185, "learning_rate": 0.0002, "epoch": 0.4587366694011485, "step": 6990}, {"loss": 0.9717, "grad_norm": 0.32561028003692627, "learning_rate": 0.0002, "epoch": 0.4593929450369155, "step": 7000}, {"loss": 0.9655, "grad_norm": 0.3388477861881256, "learning_rate": 0.0002, "epoch": 0.4600492206726825, "step": 7010}, {"loss": 0.9039, "grad_norm": 0.38045769929885864, "learning_rate": 0.0002, "epoch": 0.46070549630844954, "step": 7020}, {"loss": 0.8879, "grad_norm": 0.32500898838043213, "learning_rate": 0.0002, "epoch": 0.46136177194421657, "step": 7030}, {"loss": 0.9872, "grad_norm": 0.3758746385574341, "learning_rate": 0.0002, "epoch": 0.4620180475799836, "step": 7040}, {"loss": 0.9043, "grad_norm": 0.39033588767051697, "learning_rate": 0.0002, "epoch": 0.46267432321575064, "step": 7050}, {"loss": 0.891, "grad_norm": 0.35141658782958984, "learning_rate": 0.0002, "epoch": 0.4633305988515176, "step": 7060}, {"loss": 0.9196, "grad_norm": 0.3715546727180481, "learning_rate": 0.0002, "epoch": 0.46398687448728465, "step": 7070}, {"loss": 0.9051, "grad_norm": 0.3062947988510132, "learning_rate": 0.0002, "epoch": 0.4646431501230517, "step": 7080}, {"loss": 0.9395, "grad_norm": 0.43001696467399597, "learning_rate": 0.0002, "epoch": 0.4652994257588187, "step": 7090}, {"loss": 0.9363, "grad_norm": 0.34747597575187683, "learning_rate": 0.0002, "epoch": 0.46595570139458575, "step": 7100}, {"loss": 0.9052, "grad_norm": 0.33514827489852905, "learning_rate": 0.0002, "epoch": 0.46661197703035273, "step": 7110}, {"loss": 0.8673, "grad_norm": 0.3500545918941498, "learning_rate": 0.0002, "epoch": 0.46726825266611977, "step": 7120}, {"loss": 0.9373, "grad_norm": 0.3959280252456665, "learning_rate": 0.0002, "epoch": 0.4679245283018868, "step": 7130}, {"loss": 0.868, "grad_norm": 0.3119291067123413, "learning_rate": 0.0002, "epoch": 0.46858080393765383, "step": 7140}, {"loss": 0.923, "grad_norm": 0.36544349789619446, "learning_rate": 0.0002, "epoch": 0.4692370795734208, "step": 7150}, {"loss": 0.9281, "grad_norm": 0.3375662565231323, "learning_rate": 0.0002, "epoch": 0.46989335520918785, "step": 7160}, {"loss": 0.8948, "grad_norm": 0.48938584327697754, "learning_rate": 0.0002, "epoch": 0.4705496308449549, "step": 7170}, {"loss": 0.9016, "grad_norm": 0.36706942319869995, "learning_rate": 0.0002, "epoch": 0.4712059064807219, "step": 7180}, {"loss": 0.906, "grad_norm": 0.43472692370414734, "learning_rate": 0.0002, "epoch": 0.47186218211648895, "step": 7190}, {"loss": 0.9114, "grad_norm": 0.3433953821659088, "learning_rate": 0.0002, "epoch": 0.4725184577522559, "step": 7200}, {"loss": 0.899, "grad_norm": 0.3224952518939972, "learning_rate": 0.0002, "epoch": 0.47317473338802296, "step": 7210}, {"loss": 0.9039, "grad_norm": 0.34640923142433167, "learning_rate": 0.0002, "epoch": 0.47383100902379, "step": 7220}, {"loss": 0.9372, "grad_norm": 0.36221131682395935, "learning_rate": 0.0002, "epoch": 0.47448728465955703, "step": 7230}, {"loss": 0.9328, "grad_norm": 0.3695160150527954, "learning_rate": 0.0002, "epoch": 0.47514356029532406, "step": 7240}, {"loss": 0.8726, "grad_norm": 0.3915520906448364, "learning_rate": 0.0002, "epoch": 0.47579983593109104, "step": 7250}, {"loss": 0.9305, "grad_norm": 0.34111252427101135, "learning_rate": 0.0002, "epoch": 0.4764561115668581, "step": 7260}, {"loss": 0.8853, "grad_norm": 0.3156047761440277, "learning_rate": 0.0002, "epoch": 0.4771123872026251, "step": 7270}, {"loss": 0.9442, "grad_norm": 0.349992573261261, "learning_rate": 0.0002, "epoch": 0.47776866283839214, "step": 7280}, {"loss": 0.8841, "grad_norm": 0.3627530634403229, "learning_rate": 0.0002, "epoch": 0.4784249384741591, "step": 7290}, {"loss": 0.937, "grad_norm": 0.39342308044433594, "learning_rate": 0.0002, "epoch": 0.47908121410992616, "step": 7300}, {"loss": 0.951, "grad_norm": 0.321122944355011, "learning_rate": 0.0002, "epoch": 0.4797374897456932, "step": 7310}, {"loss": 0.9085, "grad_norm": 0.37064728140830994, "learning_rate": 0.0002, "epoch": 0.4803937653814602, "step": 7320}, {"loss": 0.9479, "grad_norm": 0.348779559135437, "learning_rate": 0.0002, "epoch": 0.48105004101722726, "step": 7330}, {"loss": 0.8651, "grad_norm": 0.3484368324279785, "learning_rate": 0.0002, "epoch": 0.48170631665299424, "step": 7340}, {"loss": 0.9096, "grad_norm": 0.3343070149421692, "learning_rate": 0.0002, "epoch": 0.48236259228876127, "step": 7350}, {"loss": 0.9148, "grad_norm": 0.3310532867908478, "learning_rate": 0.0002, "epoch": 0.4830188679245283, "step": 7360}, {"loss": 0.8652, "grad_norm": 0.37587985396385193, "learning_rate": 0.0002, "epoch": 0.48367514356029534, "step": 7370}, {"loss": 0.9302, "grad_norm": 0.37325888872146606, "learning_rate": 0.0002, "epoch": 0.4843314191960624, "step": 7380}, {"loss": 0.9382, "grad_norm": 0.3869538903236389, "learning_rate": 0.0002, "epoch": 0.48498769483182935, "step": 7390}, {"loss": 0.8967, "grad_norm": 0.34064534306526184, "learning_rate": 0.0002, "epoch": 0.4856439704675964, "step": 7400}, {"loss": 0.8884, "grad_norm": 0.35124653577804565, "learning_rate": 0.0002, "epoch": 0.4863002461033634, "step": 7410}, {"loss": 0.9178, "grad_norm": 0.38339361548423767, "learning_rate": 0.0002, "epoch": 0.48695652173913045, "step": 7420}, {"loss": 0.8984, "grad_norm": 0.32823026180267334, "learning_rate": 0.0002, "epoch": 0.48761279737489743, "step": 7430}, {"loss": 0.9348, "grad_norm": 0.32759153842926025, "learning_rate": 0.0002, "epoch": 0.48826907301066447, "step": 7440}, {"loss": 0.9023, "grad_norm": 0.3855777680873871, "learning_rate": 0.0002, "epoch": 0.4889253486464315, "step": 7450}, {"loss": 0.9201, "grad_norm": 0.34890854358673096, "learning_rate": 0.0002, "epoch": 0.48958162428219854, "step": 7460}, {"loss": 0.8947, "grad_norm": 0.3580712378025055, "learning_rate": 0.0002, "epoch": 0.49023789991796557, "step": 7470}, {"loss": 0.8413, "grad_norm": 0.34897708892822266, "learning_rate": 0.0002, "epoch": 0.49089417555373255, "step": 7480}, {"loss": 0.8508, "grad_norm": 0.32964155077934265, "learning_rate": 0.0002, "epoch": 0.4915504511894996, "step": 7490}, {"loss": 0.9298, "grad_norm": 0.3805377185344696, "learning_rate": 0.0002, "epoch": 0.4922067268252666, "step": 7500}, {"loss": 0.9354, "grad_norm": 0.3312858045101166, "learning_rate": 0.0002, "epoch": 0.49286300246103365, "step": 7510}, {"loss": 0.8966, "grad_norm": 0.32818418741226196, "learning_rate": 0.0002, "epoch": 0.49351927809680063, "step": 7520}, {"loss": 0.9036, "grad_norm": 0.3721756339073181, "learning_rate": 0.0002, "epoch": 0.49417555373256766, "step": 7530}, {"loss": 0.9483, "grad_norm": 0.4272252321243286, "learning_rate": 0.0002, "epoch": 0.4948318293683347, "step": 7540}, {"loss": 0.9159, "grad_norm": 0.3034988343715668, "learning_rate": 0.0002, "epoch": 0.49548810500410173, "step": 7550}, {"loss": 0.8996, "grad_norm": 0.34006139636039734, "learning_rate": 0.0002, "epoch": 0.49614438063986877, "step": 7560}, {"loss": 0.9122, "grad_norm": 0.33074137568473816, "learning_rate": 0.0002, "epoch": 0.49680065627563574, "step": 7570}, {"loss": 0.877, "grad_norm": 0.3747742772102356, "learning_rate": 0.0002, "epoch": 0.4974569319114028, "step": 7580}, {"loss": 0.8738, "grad_norm": 0.29682815074920654, "learning_rate": 0.0002, "epoch": 0.4981132075471698, "step": 7590}, {"loss": 0.9669, "grad_norm": 0.36150112748146057, "learning_rate": 0.0002, "epoch": 0.49876948318293685, "step": 7600}, {"loss": 0.8936, "grad_norm": 0.3893260955810547, "learning_rate": 0.0002, "epoch": 0.4994257588187039, "step": 7610}, {"loss": 0.9273, "grad_norm": 0.358791708946228, "learning_rate": 0.0002, "epoch": 0.5000820344544709, "step": 7620}, {"loss": 0.9053, "grad_norm": 0.3799469470977783, "learning_rate": 0.0002, "epoch": 0.500738310090238, "step": 7630}, {"loss": 0.9655, "grad_norm": 0.35908520221710205, "learning_rate": 0.0002, "epoch": 0.5013945857260049, "step": 7640}, {"loss": 0.8993, "grad_norm": 0.33538103103637695, "learning_rate": 0.0002, "epoch": 0.5020508613617719, "step": 7650}, {"loss": 0.8812, "grad_norm": 0.2990545928478241, "learning_rate": 0.0002, "epoch": 0.502707136997539, "step": 7660}, {"loss": 0.8965, "grad_norm": 0.3342699408531189, "learning_rate": 0.0002, "epoch": 0.503363412633306, "step": 7670}, {"loss": 0.9476, "grad_norm": 0.3656594455242157, "learning_rate": 0.0002, "epoch": 0.5040196882690731, "step": 7680}, {"loss": 0.9225, "grad_norm": 0.315176784992218, "learning_rate": 0.0002, "epoch": 0.50467596390484, "step": 7690}, {"loss": 0.8998, "grad_norm": 0.3070623576641083, "learning_rate": 0.0002, "epoch": 0.505332239540607, "step": 7700}, {"loss": 0.8565, "grad_norm": 0.3741548955440521, "learning_rate": 0.0002, "epoch": 0.5059885151763741, "step": 7710}, {"loss": 0.8882, "grad_norm": 0.3047017753124237, "learning_rate": 0.0002, "epoch": 0.5066447908121411, "step": 7720}, {"loss": 0.9002, "grad_norm": 0.5524629354476929, "learning_rate": 0.0002, "epoch": 0.5073010664479081, "step": 7730}, {"loss": 0.901, "grad_norm": 0.39234456419944763, "learning_rate": 0.0002, "epoch": 0.5079573420836752, "step": 7740}, {"loss": 0.8948, "grad_norm": 0.3306262791156769, "learning_rate": 0.0002, "epoch": 0.5086136177194421, "step": 7750}, {"loss": 0.8627, "grad_norm": 0.3181036710739136, "learning_rate": 0.0002, "epoch": 0.5092698933552092, "step": 7760}, {"loss": 0.9068, "grad_norm": 0.3531821668148041, "learning_rate": 0.0002, "epoch": 0.5099261689909762, "step": 7770}, {"loss": 0.9452, "grad_norm": 0.39781567454338074, "learning_rate": 0.0002, "epoch": 0.5105824446267432, "step": 7780}, {"loss": 0.8829, "grad_norm": 0.32413530349731445, "learning_rate": 0.0002, "epoch": 0.5112387202625103, "step": 7790}, {"loss": 0.9162, "grad_norm": 0.33857491612434387, "learning_rate": 0.0002, "epoch": 0.5118949958982772, "step": 7800}, {"loss": 0.9183, "grad_norm": 0.33426207304000854, "learning_rate": 0.0002, "epoch": 0.5125512715340443, "step": 7810}, {"loss": 0.8487, "grad_norm": 0.31170961260795593, "learning_rate": 0.0002, "epoch": 0.5132075471698113, "step": 7820}, {"loss": 0.9425, "grad_norm": 0.3401695787906647, "learning_rate": 0.0002, "epoch": 0.5138638228055783, "step": 7830}, {"loss": 0.8852, "grad_norm": 0.4325079619884491, "learning_rate": 0.0002, "epoch": 0.5145200984413454, "step": 7840}, {"loss": 0.9165, "grad_norm": 0.33266621828079224, "learning_rate": 0.0002, "epoch": 0.5151763740771124, "step": 7850}, {"loss": 0.9178, "grad_norm": 0.353476345539093, "learning_rate": 0.0002, "epoch": 0.5158326497128795, "step": 7860}, {"loss": 0.9734, "grad_norm": 0.34789302945137024, "learning_rate": 0.0002, "epoch": 0.5164889253486464, "step": 7870}, {"loss": 0.8674, "grad_norm": 0.3378570079803467, "learning_rate": 0.0002, "epoch": 0.5171452009844134, "step": 7880}, {"loss": 0.9147, "grad_norm": 0.340122789144516, "learning_rate": 0.0002, "epoch": 0.5178014766201805, "step": 7890}, {"loss": 0.9627, "grad_norm": 0.39235031604766846, "learning_rate": 0.0002, "epoch": 0.5184577522559475, "step": 7900}, {"loss": 0.9003, "grad_norm": 0.31752005219459534, "learning_rate": 0.0002, "epoch": 0.5191140278917146, "step": 7910}, {"loss": 0.9353, "grad_norm": 0.3948894739151001, "learning_rate": 0.0002, "epoch": 0.5197703035274815, "step": 7920}, {"loss": 0.8697, "grad_norm": 0.29803305864334106, "learning_rate": 0.0002, "epoch": 0.5204265791632485, "step": 7930}, {"loss": 0.9912, "grad_norm": 0.34285855293273926, "learning_rate": 0.0002, "epoch": 0.5210828547990156, "step": 7940}, {"loss": 0.9241, "grad_norm": 0.40978604555130005, "learning_rate": 0.0002, "epoch": 0.5217391304347826, "step": 7950}, {"loss": 0.9609, "grad_norm": 0.35473906993865967, "learning_rate": 0.0002, "epoch": 0.5223954060705497, "step": 7960}, {"loss": 0.925, "grad_norm": 0.3896619379520416, "learning_rate": 0.0002, "epoch": 0.5230516817063167, "step": 7970}, {"loss": 0.8863, "grad_norm": 0.3622605502605438, "learning_rate": 0.0002, "epoch": 0.5237079573420836, "step": 7980}, {"loss": 0.939, "grad_norm": 0.4097590744495392, "learning_rate": 0.0002, "epoch": 0.5243642329778507, "step": 7990}, {"loss": 0.8895, "grad_norm": 0.31397542357444763, "learning_rate": 0.0002, "epoch": 0.5250205086136177, "step": 8000}, {"loss": 0.9185, "grad_norm": 0.45559775829315186, "learning_rate": 0.0002, "epoch": 0.5256767842493847, "step": 8010}, {"loss": 0.9318, "grad_norm": 0.3471437692642212, "learning_rate": 0.0002, "epoch": 0.5263330598851518, "step": 8020}, {"loss": 0.8684, "grad_norm": 0.3105354607105255, "learning_rate": 0.0002, "epoch": 0.5269893355209188, "step": 8030}, {"loss": 0.8633, "grad_norm": 0.42847758531570435, "learning_rate": 0.0002, "epoch": 0.5276456111566858, "step": 8040}, {"loss": 0.924, "grad_norm": 0.732471227645874, "learning_rate": 0.0002, "epoch": 0.5283018867924528, "step": 8050}, {"loss": 0.8332, "grad_norm": 0.3242183327674866, "learning_rate": 0.0002, "epoch": 0.5289581624282198, "step": 8060}, {"loss": 0.9531, "grad_norm": 0.36898288130760193, "learning_rate": 0.0002, "epoch": 0.5296144380639869, "step": 8070}, {"loss": 0.933, "grad_norm": 0.3850383162498474, "learning_rate": 0.0002, "epoch": 0.5302707136997539, "step": 8080}, {"loss": 0.9184, "grad_norm": 0.3375824987888336, "learning_rate": 0.0002, "epoch": 0.530926989335521, "step": 8090}, {"loss": 0.9284, "grad_norm": 0.4321442246437073, "learning_rate": 0.0002, "epoch": 0.5315832649712879, "step": 8100}, {"loss": 0.9389, "grad_norm": 0.3836155831813812, "learning_rate": 0.0002, "epoch": 0.5322395406070549, "step": 8110}, {"loss": 0.9117, "grad_norm": 0.36536362767219543, "learning_rate": 0.0002, "epoch": 0.532895816242822, "step": 8120}, {"loss": 0.8724, "grad_norm": 0.4364359676837921, "learning_rate": 0.0002, "epoch": 0.533552091878589, "step": 8130}, {"loss": 0.8866, "grad_norm": 0.413506418466568, "learning_rate": 0.0002, "epoch": 0.5342083675143561, "step": 8140}, {"loss": 0.8985, "grad_norm": 0.33120569586753845, "learning_rate": 0.0002, "epoch": 0.534864643150123, "step": 8150}, {"loss": 0.9553, "grad_norm": 0.3408608138561249, "learning_rate": 0.0002, "epoch": 0.53552091878589, "step": 8160}, {"loss": 0.888, "grad_norm": 0.30265191197395325, "learning_rate": 0.0002, "epoch": 0.5361771944216571, "step": 8170}, {"loss": 0.8799, "grad_norm": 0.40208759903907776, "learning_rate": 0.0002, "epoch": 0.5368334700574241, "step": 8180}, {"loss": 0.9343, "grad_norm": 0.38238683342933655, "learning_rate": 0.0002, "epoch": 0.5374897456931912, "step": 8190}, {"loss": 0.9051, "grad_norm": 0.3405042290687561, "learning_rate": 0.0002, "epoch": 0.5381460213289582, "step": 8200}, {"loss": 0.9355, "grad_norm": 0.3806214928627014, "learning_rate": 0.0002, "epoch": 0.5388022969647251, "step": 8210}, {"loss": 0.9148, "grad_norm": 0.3639737069606781, "learning_rate": 0.0002, "epoch": 0.5394585726004922, "step": 8220}, {"loss": 0.9297, "grad_norm": 0.3366181552410126, "learning_rate": 0.0002, "epoch": 0.5401148482362592, "step": 8230}, {"loss": 0.9439, "grad_norm": 0.34822607040405273, "learning_rate": 0.0002, "epoch": 0.5407711238720263, "step": 8240}, {"loss": 0.925, "grad_norm": 0.3682822287082672, "learning_rate": 0.0002, "epoch": 0.5414273995077933, "step": 8250}, {"loss": 0.9246, "grad_norm": 0.39051371812820435, "learning_rate": 0.0002, "epoch": 0.5420836751435603, "step": 8260}, {"loss": 0.9008, "grad_norm": 0.37731602787971497, "learning_rate": 0.0002, "epoch": 0.5427399507793274, "step": 8270}, {"loss": 0.8849, "grad_norm": 0.34222212433815, "learning_rate": 0.0002, "epoch": 0.5433962264150943, "step": 8280}, {"loss": 0.8956, "grad_norm": 0.3870106637477875, "learning_rate": 0.0002, "epoch": 0.5440525020508613, "step": 8290}, {"loss": 0.8974, "grad_norm": 0.35894039273262024, "learning_rate": 0.0002, "epoch": 0.5447087776866284, "step": 8300}, {"loss": 0.9664, "grad_norm": 0.38450875878334045, "learning_rate": 0.0002, "epoch": 0.5453650533223954, "step": 8310}, {"loss": 0.9024, "grad_norm": 0.3106869161128998, "learning_rate": 0.0002, "epoch": 0.5460213289581625, "step": 8320}, {"loss": 0.8578, "grad_norm": 0.35935118794441223, "learning_rate": 0.0002, "epoch": 0.5466776045939294, "step": 8330}, {"loss": 0.904, "grad_norm": 0.37256261706352234, "learning_rate": 0.0002, "epoch": 0.5473338802296964, "step": 8340}, {"loss": 0.9091, "grad_norm": 0.3807767629623413, "learning_rate": 0.0002, "epoch": 0.5479901558654635, "step": 8350}, {"loss": 0.9314, "grad_norm": 0.3551439046859741, "learning_rate": 0.0002, "epoch": 0.5486464315012305, "step": 8360}, {"loss": 0.9046, "grad_norm": 0.39783644676208496, "learning_rate": 0.0002, "epoch": 0.5493027071369976, "step": 8370}, {"loss": 0.9466, "grad_norm": 0.37729522585868835, "learning_rate": 0.0002, "epoch": 0.5499589827727646, "step": 8380}, {"loss": 0.9559, "grad_norm": 0.32453545928001404, "learning_rate": 0.0002, "epoch": 0.5506152584085315, "step": 8390}, {"loss": 0.9073, "grad_norm": 0.353837251663208, "learning_rate": 0.0002, "epoch": 0.5512715340442986, "step": 8400}, {"loss": 0.961, "grad_norm": 0.37711212038993835, "learning_rate": 0.0002, "epoch": 0.5519278096800656, "step": 8410}, {"loss": 0.9844, "grad_norm": 0.3966476619243622, "learning_rate": 0.0002, "epoch": 0.5525840853158327, "step": 8420}, {"loss": 0.9221, "grad_norm": 0.338074266910553, "learning_rate": 0.0002, "epoch": 0.5532403609515997, "step": 8430}, {"loss": 0.9101, "grad_norm": 0.3984934389591217, "learning_rate": 0.0002, "epoch": 0.5538966365873667, "step": 8440}, {"loss": 0.9177, "grad_norm": 0.3430469334125519, "learning_rate": 0.0002, "epoch": 0.5545529122231337, "step": 8450}, {"loss": 0.9128, "grad_norm": 0.34223586320877075, "learning_rate": 0.0002, "epoch": 0.5552091878589007, "step": 8460}, {"loss": 0.8961, "grad_norm": 0.4057091474533081, "learning_rate": 0.0002, "epoch": 0.5558654634946678, "step": 8470}, {"loss": 0.9121, "grad_norm": 0.39272257685661316, "learning_rate": 0.0002, "epoch": 0.5565217391304348, "step": 8480}, {"loss": 0.8941, "grad_norm": 0.34801942110061646, "learning_rate": 0.0002, "epoch": 0.5571780147662018, "step": 8490}, {"loss": 0.8771, "grad_norm": 0.5177333354949951, "learning_rate": 0.0002, "epoch": 0.5578342904019689, "step": 8500}, {"loss": 0.9289, "grad_norm": 0.4139311909675598, "learning_rate": 0.0002, "epoch": 0.5584905660377358, "step": 8510}, {"loss": 0.9391, "grad_norm": 0.3526647388935089, "learning_rate": 0.0002, "epoch": 0.5591468416735029, "step": 8520}, {"loss": 0.9354, "grad_norm": 0.3475663661956787, "learning_rate": 0.0002, "epoch": 0.5598031173092699, "step": 8530}, {"loss": 0.9623, "grad_norm": 0.3338899314403534, "learning_rate": 0.0002, "epoch": 0.5604593929450369, "step": 8540}, {"loss": 0.9156, "grad_norm": 0.392212837934494, "learning_rate": 0.0002, "epoch": 0.561115668580804, "step": 8550}, {"loss": 0.9469, "grad_norm": 0.3058992028236389, "learning_rate": 0.0002, "epoch": 0.561771944216571, "step": 8560}, {"loss": 0.8465, "grad_norm": 0.40603798627853394, "learning_rate": 0.0002, "epoch": 0.5624282198523379, "step": 8570}, {"loss": 0.8962, "grad_norm": 0.37745383381843567, "learning_rate": 0.0002, "epoch": 0.563084495488105, "step": 8580}, {"loss": 0.8787, "grad_norm": 0.40886175632476807, "learning_rate": 0.0002, "epoch": 0.563740771123872, "step": 8590}, {"loss": 0.9469, "grad_norm": 0.3653600215911865, "learning_rate": 0.0002, "epoch": 0.5643970467596391, "step": 8600}, {"loss": 0.8976, "grad_norm": 0.32329508662223816, "learning_rate": 0.0002, "epoch": 0.5650533223954061, "step": 8610}, {"loss": 0.9081, "grad_norm": 0.3535236716270447, "learning_rate": 0.0002, "epoch": 0.565709598031173, "step": 8620}, {"loss": 0.8703, "grad_norm": 0.37497273087501526, "learning_rate": 0.0002, "epoch": 0.5663658736669401, "step": 8630}, {"loss": 0.8853, "grad_norm": 0.3781719505786896, "learning_rate": 0.0002, "epoch": 0.5670221493027071, "step": 8640}, {"loss": 0.9074, "grad_norm": 0.4481894075870514, "learning_rate": 0.0002, "epoch": 0.5676784249384742, "step": 8650}, {"loss": 0.8923, "grad_norm": 0.35790109634399414, "learning_rate": 0.0002, "epoch": 0.5683347005742412, "step": 8660}, {"loss": 0.8946, "grad_norm": 0.4617280662059784, "learning_rate": 0.0002, "epoch": 0.5689909762100082, "step": 8670}, {"loss": 0.9059, "grad_norm": 0.3813382089138031, "learning_rate": 0.0002, "epoch": 0.5696472518457752, "step": 8680}, {"loss": 0.9126, "grad_norm": 0.33388257026672363, "learning_rate": 0.0002, "epoch": 0.5703035274815422, "step": 8690}, {"loss": 0.8866, "grad_norm": 0.330185204744339, "learning_rate": 0.0002, "epoch": 0.5709598031173093, "step": 8700}, {"loss": 0.916, "grad_norm": 0.3768845200538635, "learning_rate": 0.0002, "epoch": 0.5716160787530763, "step": 8710}, {"loss": 0.9213, "grad_norm": 0.3221369683742523, "learning_rate": 0.0002, "epoch": 0.5722723543888433, "step": 8720}, {"loss": 0.8919, "grad_norm": 0.3221580684185028, "learning_rate": 0.0002, "epoch": 0.5729286300246104, "step": 8730}, {"loss": 0.924, "grad_norm": 0.3693002760410309, "learning_rate": 0.0002, "epoch": 0.5735849056603773, "step": 8740}, {"loss": 0.8983, "grad_norm": 0.3859325349330902, "learning_rate": 0.0002, "epoch": 0.5742411812961444, "step": 8750}, {"loss": 0.9321, "grad_norm": 0.3891449570655823, "learning_rate": 0.0002, "epoch": 0.5748974569319114, "step": 8760}, {"loss": 0.9315, "grad_norm": 0.3313274681568146, "learning_rate": 0.0002, "epoch": 0.5755537325676784, "step": 8770}, {"loss": 0.89, "grad_norm": 0.33418914675712585, "learning_rate": 0.0002, "epoch": 0.5762100082034455, "step": 8780}, {"loss": 0.9029, "grad_norm": 0.4153108596801758, "learning_rate": 0.0002, "epoch": 0.5768662838392125, "step": 8790}, {"loss": 0.9126, "grad_norm": 0.3614528477191925, "learning_rate": 0.0002, "epoch": 0.5775225594749795, "step": 8800}, {"loss": 0.9086, "grad_norm": 0.3172459602355957, "learning_rate": 0.0002, "epoch": 0.5781788351107465, "step": 8810}, {"loss": 0.9225, "grad_norm": 0.3795868456363678, "learning_rate": 0.0002, "epoch": 0.5788351107465135, "step": 8820}, {"loss": 0.9199, "grad_norm": 0.3750050961971283, "learning_rate": 0.0002, "epoch": 0.5794913863822806, "step": 8830}, {"loss": 0.9527, "grad_norm": 0.3348597288131714, "learning_rate": 0.0002, "epoch": 0.5801476620180476, "step": 8840}, {"loss": 0.9379, "grad_norm": 0.39116451144218445, "learning_rate": 0.0002, "epoch": 0.5808039376538146, "step": 8850}, {"loss": 0.9087, "grad_norm": 0.33639633655548096, "learning_rate": 0.0002, "epoch": 0.5814602132895816, "step": 8860}, {"loss": 0.9181, "grad_norm": 0.33336129784584045, "learning_rate": 0.0002, "epoch": 0.5821164889253486, "step": 8870}, {"loss": 0.9194, "grad_norm": 0.3624817132949829, "learning_rate": 0.0002, "epoch": 0.5827727645611157, "step": 8880}, {"loss": 0.9059, "grad_norm": 0.2976662218570709, "learning_rate": 0.0002, "epoch": 0.5834290401968827, "step": 8890}, {"loss": 0.8903, "grad_norm": 0.4231838583946228, "learning_rate": 0.0002, "epoch": 0.5840853158326497, "step": 8900}, {"loss": 0.9157, "grad_norm": 0.343156099319458, "learning_rate": 0.0002, "epoch": 0.5847415914684168, "step": 8910}, {"loss": 0.9655, "grad_norm": 0.39717403054237366, "learning_rate": 0.0002, "epoch": 0.5853978671041837, "step": 8920}, {"loss": 0.8758, "grad_norm": 0.3388144075870514, "learning_rate": 0.0002, "epoch": 0.5860541427399508, "step": 8930}, {"loss": 0.922, "grad_norm": 0.3949063718318939, "learning_rate": 0.0002, "epoch": 0.5867104183757178, "step": 8940}, {"loss": 0.9081, "grad_norm": 0.358415424823761, "learning_rate": 0.0002, "epoch": 0.5873666940114848, "step": 8950}, {"loss": 0.906, "grad_norm": 0.367883563041687, "learning_rate": 0.0002, "epoch": 0.5880229696472519, "step": 8960}, {"loss": 0.9131, "grad_norm": 0.3859010636806488, "learning_rate": 0.0002, "epoch": 0.5886792452830188, "step": 8970}, {"loss": 0.916, "grad_norm": 0.3540095090866089, "learning_rate": 0.0002, "epoch": 0.5893355209187859, "step": 8980}, {"loss": 0.9704, "grad_norm": 0.33913302421569824, "learning_rate": 0.0002, "epoch": 0.5899917965545529, "step": 8990}, {"loss": 0.8772, "grad_norm": 0.35641804337501526, "learning_rate": 0.0002, "epoch": 0.5906480721903199, "step": 9000}, {"loss": 0.8644, "grad_norm": 0.3448907434940338, "learning_rate": 0.0002, "epoch": 0.591304347826087, "step": 9010}, {"loss": 0.9059, "grad_norm": 0.4147050082683563, "learning_rate": 0.0002, "epoch": 0.591960623461854, "step": 9020}, {"loss": 0.9134, "grad_norm": 0.45032307505607605, "learning_rate": 0.0002, "epoch": 0.592616899097621, "step": 9030}, {"loss": 0.8855, "grad_norm": 0.3628511428833008, "learning_rate": 0.0002, "epoch": 0.593273174733388, "step": 9040}, {"loss": 0.8924, "grad_norm": 0.42760607600212097, "learning_rate": 0.0002, "epoch": 0.593929450369155, "step": 9050}, {"loss": 0.889, "grad_norm": 0.3591140806674957, "learning_rate": 0.0002, "epoch": 0.5945857260049221, "step": 9060}, {"loss": 0.9409, "grad_norm": 0.3496510982513428, "learning_rate": 0.0002, "epoch": 0.5952420016406891, "step": 9070}, {"loss": 0.9394, "grad_norm": 0.4098506569862366, "learning_rate": 0.0002, "epoch": 0.5958982772764561, "step": 9080}, {"loss": 0.9455, "grad_norm": 0.3535410165786743, "learning_rate": 0.0002, "epoch": 0.5965545529122231, "step": 9090}, {"loss": 0.8814, "grad_norm": 0.3716369569301605, "learning_rate": 0.0002, "epoch": 0.5972108285479901, "step": 9100}, {"loss": 0.8656, "grad_norm": 0.3236614763736725, "learning_rate": 0.0002, "epoch": 0.5978671041837572, "step": 9110}, {"loss": 0.905, "grad_norm": 0.311577171087265, "learning_rate": 0.0002, "epoch": 0.5985233798195242, "step": 9120}, {"loss": 0.9118, "grad_norm": 0.3261461555957794, "learning_rate": 0.0002, "epoch": 0.5991796554552912, "step": 9130}, {"loss": 0.9274, "grad_norm": 0.3794492781162262, "learning_rate": 0.0002, "epoch": 0.5998359310910583, "step": 9140}, {"loss": 0.8967, "grad_norm": 0.31726357340812683, "learning_rate": 0.0002, "epoch": 0.6004922067268252, "step": 9150}, {"loss": 0.9156, "grad_norm": 0.38432174921035767, "learning_rate": 0.0002, "epoch": 0.6011484823625923, "step": 9160}, {"loss": 0.9703, "grad_norm": 0.4364495873451233, "learning_rate": 0.0002, "epoch": 0.6018047579983593, "step": 9170}, {"loss": 0.9162, "grad_norm": 0.41058987379074097, "learning_rate": 0.0002, "epoch": 0.6024610336341263, "step": 9180}, {"loss": 0.8914, "grad_norm": 0.3146302103996277, "learning_rate": 0.0002, "epoch": 0.6031173092698934, "step": 9190}, {"loss": 0.8859, "grad_norm": 0.3117610216140747, "learning_rate": 0.0002, "epoch": 0.6037735849056604, "step": 9200}, {"loss": 0.8741, "grad_norm": 0.31921523809432983, "learning_rate": 0.0002, "epoch": 0.6044298605414274, "step": 9210}, {"loss": 0.8756, "grad_norm": 0.3295772969722748, "learning_rate": 0.0002, "epoch": 0.6050861361771944, "step": 9220}, {"loss": 0.8578, "grad_norm": 0.3216910660266876, "learning_rate": 0.0002, "epoch": 0.6057424118129614, "step": 9230}, {"loss": 0.9341, "grad_norm": 0.40590721368789673, "learning_rate": 0.0002, "epoch": 0.6063986874487285, "step": 9240}, {"loss": 0.935, "grad_norm": 0.4337029457092285, "learning_rate": 0.0002, "epoch": 0.6070549630844955, "step": 9250}, {"loss": 0.8866, "grad_norm": 0.3434010148048401, "learning_rate": 0.0002, "epoch": 0.6077112387202626, "step": 9260}, {"loss": 0.8692, "grad_norm": 0.340589702129364, "learning_rate": 0.0002, "epoch": 0.6083675143560295, "step": 9270}, {"loss": 0.8741, "grad_norm": 0.33650949597358704, "learning_rate": 0.0002, "epoch": 0.6090237899917965, "step": 9280}, {"loss": 0.9396, "grad_norm": 0.4195605218410492, "learning_rate": 0.0002, "epoch": 0.6096800656275636, "step": 9290}, {"loss": 0.8877, "grad_norm": 0.444624125957489, "learning_rate": 0.0002, "epoch": 0.6103363412633306, "step": 9300}, {"loss": 0.8792, "grad_norm": 0.5425066947937012, "learning_rate": 0.0002, "epoch": 0.6109926168990977, "step": 9310}, {"loss": 0.865, "grad_norm": 0.36225831508636475, "learning_rate": 0.0002, "epoch": 0.6116488925348647, "step": 9320}, {"loss": 0.9439, "grad_norm": 0.3421785533428192, "learning_rate": 0.0002, "epoch": 0.6123051681706316, "step": 9330}, {"loss": 0.8842, "grad_norm": 0.3792729079723358, "learning_rate": 0.0002, "epoch": 0.6129614438063987, "step": 9340}, {"loss": 0.8791, "grad_norm": 0.41844120621681213, "learning_rate": 0.0002, "epoch": 0.6136177194421657, "step": 9350}, {"loss": 0.9091, "grad_norm": 0.3430991768836975, "learning_rate": 0.0002, "epoch": 0.6142739950779327, "step": 9360}, {"loss": 0.9165, "grad_norm": 0.3460402190685272, "learning_rate": 0.0002, "epoch": 0.6149302707136998, "step": 9370}, {"loss": 0.8571, "grad_norm": 0.3775254189968109, "learning_rate": 0.0002, "epoch": 0.6155865463494667, "step": 9380}, {"loss": 0.8915, "grad_norm": 0.3706645965576172, "learning_rate": 0.0002, "epoch": 0.6162428219852338, "step": 9390}, {"loss": 0.9085, "grad_norm": 0.3817055821418762, "learning_rate": 0.0002, "epoch": 0.6168990976210008, "step": 9400}, {"loss": 0.9086, "grad_norm": 0.3856641352176666, "learning_rate": 0.0002, "epoch": 0.6175553732567678, "step": 9410}, {"loss": 0.9031, "grad_norm": 0.3271346688270569, "learning_rate": 0.0002, "epoch": 0.6182116488925349, "step": 9420}, {"loss": 0.883, "grad_norm": 0.334314227104187, "learning_rate": 0.0002, "epoch": 0.6188679245283019, "step": 9430}, {"loss": 0.9398, "grad_norm": 0.3850700557231903, "learning_rate": 0.0002, "epoch": 0.619524200164069, "step": 9440}, {"loss": 0.9246, "grad_norm": 0.37492436170578003, "learning_rate": 0.0002, "epoch": 0.6201804757998359, "step": 9450}, {"loss": 0.9308, "grad_norm": 0.44262826442718506, "learning_rate": 0.0002, "epoch": 0.6208367514356029, "step": 9460}, {"loss": 0.9659, "grad_norm": 0.4070657193660736, "learning_rate": 0.0002, "epoch": 0.62149302707137, "step": 9470}, {"loss": 0.8812, "grad_norm": 0.37636154890060425, "learning_rate": 0.0002, "epoch": 0.622149302707137, "step": 9480}, {"loss": 0.941, "grad_norm": 0.38988572359085083, "learning_rate": 0.0002, "epoch": 0.6228055783429041, "step": 9490}, {"loss": 0.8959, "grad_norm": 0.36479735374450684, "learning_rate": 0.0002, "epoch": 0.623461853978671, "step": 9500}, {"loss": 0.9431, "grad_norm": 0.3759172260761261, "learning_rate": 0.0002, "epoch": 0.624118129614438, "step": 9510}, {"loss": 0.889, "grad_norm": 0.3380950689315796, "learning_rate": 0.0002, "epoch": 0.6247744052502051, "step": 9520}, {"loss": 0.8597, "grad_norm": 0.3483046889305115, "learning_rate": 0.0002, "epoch": 0.6254306808859721, "step": 9530}, {"loss": 0.9195, "grad_norm": 0.3562379777431488, "learning_rate": 0.0002, "epoch": 0.6260869565217392, "step": 9540}, {"loss": 0.9434, "grad_norm": 0.40167364478111267, "learning_rate": 0.0002, "epoch": 0.6267432321575062, "step": 9550}, {"loss": 0.8677, "grad_norm": 0.36348867416381836, "learning_rate": 0.0002, "epoch": 0.6273995077932731, "step": 9560}, {"loss": 0.9061, "grad_norm": 0.33701515197753906, "learning_rate": 0.0002, "epoch": 0.6280557834290402, "step": 9570}, {"loss": 0.8926, "grad_norm": 0.3795888125896454, "learning_rate": 0.0002, "epoch": 0.6287120590648072, "step": 9580}, {"loss": 0.8672, "grad_norm": 0.3701418340206146, "learning_rate": 0.0002, "epoch": 0.6293683347005743, "step": 9590}, {"loss": 0.9025, "grad_norm": 0.32559722661972046, "learning_rate": 0.0002, "epoch": 0.6300246103363413, "step": 9600}, {"loss": 0.9223, "grad_norm": 0.3093271553516388, "learning_rate": 0.0002, "epoch": 0.6306808859721083, "step": 9610}, {"loss": 0.9032, "grad_norm": 0.41225478053092957, "learning_rate": 0.0002, "epoch": 0.6313371616078753, "step": 9620}, {"loss": 0.9125, "grad_norm": 0.3798231780529022, "learning_rate": 0.0002, "epoch": 0.6319934372436423, "step": 9630}, {"loss": 0.9252, "grad_norm": 0.38690295815467834, "learning_rate": 0.0002, "epoch": 0.6326497128794093, "step": 9640}, {"loss": 0.8668, "grad_norm": 0.3629007637500763, "learning_rate": 0.0002, "epoch": 0.6333059885151764, "step": 9650}, {"loss": 0.9143, "grad_norm": 0.39467132091522217, "learning_rate": 0.0002, "epoch": 0.6339622641509434, "step": 9660}, {"loss": 0.9267, "grad_norm": 0.37746182084083557, "learning_rate": 0.0002, "epoch": 0.6346185397867105, "step": 9670}, {"loss": 0.8915, "grad_norm": 0.3718436062335968, "learning_rate": 0.0002, "epoch": 0.6352748154224774, "step": 9680}, {"loss": 0.9111, "grad_norm": 0.3951144218444824, "learning_rate": 0.0002, "epoch": 0.6359310910582444, "step": 9690}, {"loss": 0.897, "grad_norm": 0.4104543924331665, "learning_rate": 0.0002, "epoch": 0.6365873666940115, "step": 9700}, {"loss": 0.9349, "grad_norm": 0.37318137288093567, "learning_rate": 0.0002, "epoch": 0.6372436423297785, "step": 9710}, {"loss": 0.9009, "grad_norm": 0.35614442825317383, "learning_rate": 0.0002, "epoch": 0.6378999179655456, "step": 9720}, {"loss": 0.9858, "grad_norm": 0.37922942638397217, "learning_rate": 0.0002, "epoch": 0.6385561936013125, "step": 9730}, {"loss": 0.8533, "grad_norm": 0.35624784231185913, "learning_rate": 0.0002, "epoch": 0.6392124692370795, "step": 9740}, {"loss": 0.934, "grad_norm": 0.34946876764297485, "learning_rate": 0.0002, "epoch": 0.6398687448728466, "step": 9750}, {"loss": 0.9176, "grad_norm": 0.37666648626327515, "learning_rate": 0.0002, "epoch": 0.6405250205086136, "step": 9760}, {"loss": 0.9297, "grad_norm": 0.3619046211242676, "learning_rate": 0.0002, "epoch": 0.6411812961443807, "step": 9770}, {"loss": 0.9135, "grad_norm": 0.3814936578273773, "learning_rate": 0.0002, "epoch": 0.6418375717801477, "step": 9780}, {"loss": 0.9143, "grad_norm": 0.4181577265262604, "learning_rate": 0.0002, "epoch": 0.6424938474159146, "step": 9790}, {"loss": 0.8978, "grad_norm": 0.3707144558429718, "learning_rate": 0.0002, "epoch": 0.6431501230516817, "step": 9800}, {"loss": 0.9311, "grad_norm": 0.3364716172218323, "learning_rate": 0.0002, "epoch": 0.6438063986874487, "step": 9810}, {"loss": 0.8566, "grad_norm": 0.31425684690475464, "learning_rate": 0.0002, "epoch": 0.6444626743232158, "step": 9820}, {"loss": 0.8792, "grad_norm": 0.32928282022476196, "learning_rate": 0.0002, "epoch": 0.6451189499589828, "step": 9830}, {"loss": 0.9232, "grad_norm": 0.39178264141082764, "learning_rate": 0.0002, "epoch": 0.6457752255947498, "step": 9840}, {"loss": 0.9413, "grad_norm": 0.37753361463546753, "learning_rate": 0.0002, "epoch": 0.6464315012305168, "step": 9850}, {"loss": 0.8777, "grad_norm": 0.4574730396270752, "learning_rate": 0.0002, "epoch": 0.6470877768662838, "step": 9860}, {"loss": 0.9078, "grad_norm": 0.3926962614059448, "learning_rate": 0.0002, "epoch": 0.6477440525020509, "step": 9870}, {"loss": 0.9221, "grad_norm": 0.3919081687927246, "learning_rate": 0.0002, "epoch": 0.6484003281378179, "step": 9880}, {"loss": 0.8681, "grad_norm": 0.3904387652873993, "learning_rate": 0.0002, "epoch": 0.6490566037735849, "step": 9890}, {"loss": 0.883, "grad_norm": 0.40808236598968506, "learning_rate": 0.0002, "epoch": 0.649712879409352, "step": 9900}, {"loss": 0.9133, "grad_norm": 0.40531080961227417, "learning_rate": 0.0002, "epoch": 0.6503691550451189, "step": 9910}, {"loss": 0.9417, "grad_norm": 0.3374323844909668, "learning_rate": 0.0002, "epoch": 0.6510254306808859, "step": 9920}, {"loss": 0.9769, "grad_norm": 0.4000678062438965, "learning_rate": 0.0002, "epoch": 0.651681706316653, "step": 9930}, {"loss": 0.9069, "grad_norm": 0.35415270924568176, "learning_rate": 0.0002, "epoch": 0.65233798195242, "step": 9940}, {"loss": 0.9337, "grad_norm": 0.3612231910228729, "learning_rate": 0.0002, "epoch": 0.6529942575881871, "step": 9950}, {"loss": 0.9402, "grad_norm": 0.3081146776676178, "learning_rate": 0.0002, "epoch": 0.6536505332239541, "step": 9960}, {"loss": 0.8942, "grad_norm": 0.31079018115997314, "learning_rate": 0.0002, "epoch": 0.654306808859721, "step": 9970}, {"loss": 0.905, "grad_norm": 0.4256346523761749, "learning_rate": 0.0002, "epoch": 0.6549630844954881, "step": 9980}, {"loss": 0.9226, "grad_norm": 0.3592916429042816, "learning_rate": 0.0002, "epoch": 0.6556193601312551, "step": 9990}, {"loss": 0.9018, "grad_norm": 0.3496004641056061, "learning_rate": 0.0002, "epoch": 0.6562756357670222, "step": 10000}, {"loss": 0.931, "grad_norm": 0.4177037179470062, "learning_rate": 0.0002, "epoch": 0.6569319114027892, "step": 10010}, {"loss": 0.9034, "grad_norm": 0.36680638790130615, "learning_rate": 0.0002, "epoch": 0.6575881870385561, "step": 10020}, {"loss": 0.9706, "grad_norm": 0.3916943073272705, "learning_rate": 0.0002, "epoch": 0.6582444626743232, "step": 10030}, {"loss": 0.886, "grad_norm": 0.44537290930747986, "learning_rate": 0.0002, "epoch": 0.6589007383100902, "step": 10040}, {"loss": 0.8798, "grad_norm": 0.3404697775840759, "learning_rate": 0.0002, "epoch": 0.6595570139458573, "step": 10050}, {"loss": 0.9801, "grad_norm": 0.3304594159126282, "learning_rate": 0.0002, "epoch": 0.6602132895816243, "step": 10060}, {"loss": 0.9107, "grad_norm": 0.39067313075065613, "learning_rate": 0.0002, "epoch": 0.6608695652173913, "step": 10070}, {"loss": 0.9336, "grad_norm": 0.366178959608078, "learning_rate": 0.0002, "epoch": 0.6615258408531584, "step": 10080}, {"loss": 0.8817, "grad_norm": 0.3880734145641327, "learning_rate": 0.0002, "epoch": 0.6621821164889253, "step": 10090}, {"loss": 0.9087, "grad_norm": 0.42047396302223206, "learning_rate": 0.0002, "epoch": 0.6628383921246924, "step": 10100}, {"loss": 0.8859, "grad_norm": 0.3299349546432495, "learning_rate": 0.0002, "epoch": 0.6634946677604594, "step": 10110}, {"loss": 0.9461, "grad_norm": 0.33470937609672546, "learning_rate": 0.0002, "epoch": 0.6641509433962264, "step": 10120}, {"loss": 0.9453, "grad_norm": 0.36240577697753906, "learning_rate": 0.0002, "epoch": 0.6648072190319935, "step": 10130}, {"loss": 0.8706, "grad_norm": 0.41457104682922363, "learning_rate": 0.0002, "epoch": 0.6654634946677604, "step": 10140}, {"loss": 0.9385, "grad_norm": 0.39064788818359375, "learning_rate": 0.0002, "epoch": 0.6661197703035275, "step": 10150}, {"loss": 0.9252, "grad_norm": 0.4137183725833893, "learning_rate": 0.0002, "epoch": 0.6667760459392945, "step": 10160}, {"loss": 0.8874, "grad_norm": 0.363413542509079, "learning_rate": 0.0002, "epoch": 0.6674323215750615, "step": 10170}, {"loss": 0.8936, "grad_norm": 0.3631424307823181, "learning_rate": 0.0002, "epoch": 0.6680885972108286, "step": 10180}, {"loss": 0.9181, "grad_norm": 0.460721880197525, "learning_rate": 0.0002, "epoch": 0.6687448728465956, "step": 10190}, {"loss": 0.8879, "grad_norm": 0.3525084853172302, "learning_rate": 0.0002, "epoch": 0.6694011484823625, "step": 10200}, {"loss": 0.934, "grad_norm": 0.34321045875549316, "learning_rate": 0.0002, "epoch": 0.6700574241181296, "step": 10210}, {"loss": 0.9056, "grad_norm": 0.3397759795188904, "learning_rate": 0.0002, "epoch": 0.6707136997538966, "step": 10220}, {"loss": 0.8883, "grad_norm": 0.3680257201194763, "learning_rate": 0.0002, "epoch": 0.6713699753896637, "step": 10230}, {"loss": 0.9316, "grad_norm": 0.4023214876651764, "learning_rate": 0.0002, "epoch": 0.6720262510254307, "step": 10240}, {"loss": 0.8809, "grad_norm": 0.3645709753036499, "learning_rate": 0.0002, "epoch": 0.6726825266611977, "step": 10250}, {"loss": 0.9766, "grad_norm": 0.3558615744113922, "learning_rate": 0.0002, "epoch": 0.6733388022969647, "step": 10260}, {"loss": 0.872, "grad_norm": 0.3725513517856598, "learning_rate": 0.0002, "epoch": 0.6739950779327317, "step": 10270}, {"loss": 0.9354, "grad_norm": 0.42790961265563965, "learning_rate": 0.0002, "epoch": 0.6746513535684988, "step": 10280}, {"loss": 0.8955, "grad_norm": 0.3808377683162689, "learning_rate": 0.0002, "epoch": 0.6753076292042658, "step": 10290}, {"loss": 0.8761, "grad_norm": 0.40500468015670776, "learning_rate": 0.0002, "epoch": 0.6759639048400328, "step": 10300}, {"loss": 0.9212, "grad_norm": 0.3631184995174408, "learning_rate": 0.0002, "epoch": 0.6766201804757999, "step": 10310}, {"loss": 0.8867, "grad_norm": 0.3319573998451233, "learning_rate": 0.0002, "epoch": 0.6772764561115668, "step": 10320}, {"loss": 0.857, "grad_norm": 0.3851188123226166, "learning_rate": 0.0002, "epoch": 0.6779327317473339, "step": 10330}, {"loss": 0.9159, "grad_norm": 0.37413159012794495, "learning_rate": 0.0002, "epoch": 0.6785890073831009, "step": 10340}, {"loss": 0.8728, "grad_norm": 0.34124433994293213, "learning_rate": 0.0002, "epoch": 0.6792452830188679, "step": 10350}, {"loss": 0.9654, "grad_norm": 0.3786381781101227, "learning_rate": 0.0002, "epoch": 0.679901558654635, "step": 10360}, {"loss": 0.915, "grad_norm": 0.3848523199558258, "learning_rate": 0.0002, "epoch": 0.680557834290402, "step": 10370}, {"loss": 0.8873, "grad_norm": 0.29365262389183044, "learning_rate": 0.0002, "epoch": 0.681214109926169, "step": 10380}, {"loss": 0.8692, "grad_norm": 0.3481557369232178, "learning_rate": 0.0002, "epoch": 0.681870385561936, "step": 10390}, {"loss": 0.885, "grad_norm": 0.38596320152282715, "learning_rate": 0.0002, "epoch": 0.682526661197703, "step": 10400}, {"loss": 0.8545, "grad_norm": 0.30872032046318054, "learning_rate": 0.0002, "epoch": 0.6831829368334701, "step": 10410}, {"loss": 0.867, "grad_norm": 0.3729351758956909, "learning_rate": 0.0002, "epoch": 0.6838392124692371, "step": 10420}, {"loss": 0.9089, "grad_norm": 0.35829052329063416, "learning_rate": 0.0002, "epoch": 0.684495488105004, "step": 10430}, {"loss": 0.8797, "grad_norm": 0.34916967153549194, "learning_rate": 0.0002, "epoch": 0.6851517637407711, "step": 10440}, {"loss": 0.8731, "grad_norm": 0.40377020835876465, "learning_rate": 0.0002, "epoch": 0.6858080393765381, "step": 10450}, {"loss": 0.9027, "grad_norm": 0.36841881275177, "learning_rate": 0.0002, "epoch": 0.6864643150123052, "step": 10460}, {"loss": 0.8808, "grad_norm": 0.35462167859077454, "learning_rate": 0.0002, "epoch": 0.6871205906480722, "step": 10470}, {"loss": 0.967, "grad_norm": 0.4630918502807617, "learning_rate": 0.0002, "epoch": 0.6877768662838392, "step": 10480}, {"loss": 0.9114, "grad_norm": 0.4230295419692993, "learning_rate": 0.0002, "epoch": 0.6884331419196063, "step": 10490}, {"loss": 0.8908, "grad_norm": 0.35059425234794617, "learning_rate": 0.0002, "epoch": 0.6890894175553732, "step": 10500}, {"loss": 0.9325, "grad_norm": 0.4072548747062683, "learning_rate": 0.0002, "epoch": 0.6897456931911403, "step": 10510}, {"loss": 0.9026, "grad_norm": 0.33755314350128174, "learning_rate": 0.0002, "epoch": 0.6904019688269073, "step": 10520}, {"loss": 0.908, "grad_norm": 0.4411669075489044, "learning_rate": 0.0002, "epoch": 0.6910582444626743, "step": 10530}, {"loss": 0.8639, "grad_norm": 0.9155740737915039, "learning_rate": 0.0002, "epoch": 0.6917145200984414, "step": 10540}, {"loss": 0.8975, "grad_norm": 0.39329391717910767, "learning_rate": 0.0002, "epoch": 0.6923707957342083, "step": 10550}, {"loss": 0.92, "grad_norm": 0.3602689206600189, "learning_rate": 0.0002, "epoch": 0.6930270713699754, "step": 10560}, {"loss": 0.8247, "grad_norm": 0.35848474502563477, "learning_rate": 0.0002, "epoch": 0.6936833470057424, "step": 10570}, {"loss": 0.9171, "grad_norm": 0.37857621908187866, "learning_rate": 0.0002, "epoch": 0.6943396226415094, "step": 10580}, {"loss": 0.9533, "grad_norm": 0.37071332335472107, "learning_rate": 0.0002, "epoch": 0.6949958982772765, "step": 10590}, {"loss": 0.9237, "grad_norm": 0.33589425683021545, "learning_rate": 0.0002, "epoch": 0.6956521739130435, "step": 10600}, {"loss": 0.8772, "grad_norm": 0.4279285669326782, "learning_rate": 0.0002, "epoch": 0.6963084495488105, "step": 10610}, {"loss": 0.9452, "grad_norm": 0.47047463059425354, "learning_rate": 0.0002, "epoch": 0.6969647251845775, "step": 10620}, {"loss": 0.8996, "grad_norm": 0.41214805841445923, "learning_rate": 0.0002, "epoch": 0.6976210008203445, "step": 10630}, {"loss": 0.8776, "grad_norm": 0.3999953866004944, "learning_rate": 0.0002, "epoch": 0.6982772764561116, "step": 10640}, {"loss": 0.9426, "grad_norm": 0.42441290616989136, "learning_rate": 0.0002, "epoch": 0.6989335520918786, "step": 10650}, {"loss": 0.9319, "grad_norm": 0.41990748047828674, "learning_rate": 0.0002, "epoch": 0.6995898277276457, "step": 10660}, {"loss": 0.9024, "grad_norm": 0.37445810437202454, "learning_rate": 0.0002, "epoch": 0.7002461033634126, "step": 10670}, {"loss": 0.9029, "grad_norm": 0.37638649344444275, "learning_rate": 0.0002, "epoch": 0.7009023789991796, "step": 10680}, {"loss": 0.9211, "grad_norm": 0.41475534439086914, "learning_rate": 0.0002, "epoch": 0.7015586546349467, "step": 10690}, {"loss": 0.914, "grad_norm": 0.3430996537208557, "learning_rate": 0.0002, "epoch": 0.7022149302707137, "step": 10700}, {"loss": 0.9051, "grad_norm": 0.3569522798061371, "learning_rate": 0.0002, "epoch": 0.7028712059064807, "step": 10710}, {"loss": 0.8792, "grad_norm": 0.3145142197608948, "learning_rate": 0.0002, "epoch": 0.7035274815422478, "step": 10720}, {"loss": 0.9537, "grad_norm": 0.40700432658195496, "learning_rate": 0.0002, "epoch": 0.7041837571780147, "step": 10730}, {"loss": 0.8677, "grad_norm": 0.37982651591300964, "learning_rate": 0.0002, "epoch": 0.7048400328137818, "step": 10740}, {"loss": 0.8899, "grad_norm": 0.39240679144859314, "learning_rate": 0.0002, "epoch": 0.7054963084495488, "step": 10750}, {"loss": 0.9516, "grad_norm": 0.41788724064826965, "learning_rate": 0.0002, "epoch": 0.7061525840853158, "step": 10760}, {"loss": 0.8672, "grad_norm": 0.4015905559062958, "learning_rate": 0.0002, "epoch": 0.7068088597210829, "step": 10770}, {"loss": 0.9131, "grad_norm": 0.4019724130630493, "learning_rate": 0.0002, "epoch": 0.7074651353568499, "step": 10780}, {"loss": 0.8698, "grad_norm": 0.37387898564338684, "learning_rate": 0.0002, "epoch": 0.7081214109926169, "step": 10790}, {"loss": 0.8547, "grad_norm": 0.32087528705596924, "learning_rate": 0.0002, "epoch": 0.7087776866283839, "step": 10800}, {"loss": 0.9241, "grad_norm": 0.38160258531570435, "learning_rate": 0.0002, "epoch": 0.7094339622641509, "step": 10810}, {"loss": 0.8941, "grad_norm": 0.340973436832428, "learning_rate": 0.0002, "epoch": 0.710090237899918, "step": 10820}, {"loss": 0.9404, "grad_norm": 0.3848867416381836, "learning_rate": 0.0002, "epoch": 0.710746513535685, "step": 10830}, {"loss": 0.9299, "grad_norm": 0.36736220121383667, "learning_rate": 0.0002, "epoch": 0.711402789171452, "step": 10840}, {"loss": 0.8881, "grad_norm": 0.367404967546463, "learning_rate": 0.0002, "epoch": 0.712059064807219, "step": 10850}, {"loss": 0.9224, "grad_norm": 0.333751380443573, "learning_rate": 0.0002, "epoch": 0.712715340442986, "step": 10860}, {"loss": 0.8712, "grad_norm": 0.3865894079208374, "learning_rate": 0.0002, "epoch": 0.7133716160787531, "step": 10870}, {"loss": 0.8957, "grad_norm": 0.34016433358192444, "learning_rate": 0.0002, "epoch": 0.7140278917145201, "step": 10880}, {"loss": 0.8493, "grad_norm": 0.3233864903450012, "learning_rate": 0.0002, "epoch": 0.7146841673502872, "step": 10890}, {"loss": 0.9148, "grad_norm": 0.4275553822517395, "learning_rate": 0.0002, "epoch": 0.7153404429860541, "step": 10900}, {"loss": 0.9345, "grad_norm": 0.41715168952941895, "learning_rate": 0.0002, "epoch": 0.7159967186218211, "step": 10910}, {"loss": 0.9321, "grad_norm": 0.40540871024131775, "learning_rate": 0.0002, "epoch": 0.7166529942575882, "step": 10920}, {"loss": 0.9035, "grad_norm": 0.35520824790000916, "learning_rate": 0.0002, "epoch": 0.7173092698933552, "step": 10930}, {"loss": 0.9243, "grad_norm": 0.3805985748767853, "learning_rate": 0.0002, "epoch": 0.7179655455291223, "step": 10940}, {"loss": 0.9542, "grad_norm": 0.36969226598739624, "learning_rate": 0.0002, "epoch": 0.7186218211648893, "step": 10950}, {"loss": 0.892, "grad_norm": 0.38408684730529785, "learning_rate": 0.0002, "epoch": 0.7192780968006562, "step": 10960}, {"loss": 0.8699, "grad_norm": 0.3865699768066406, "learning_rate": 0.0002, "epoch": 0.7199343724364233, "step": 10970}, {"loss": 0.8767, "grad_norm": 0.37172383069992065, "learning_rate": 0.0002, "epoch": 0.7205906480721903, "step": 10980}, {"loss": 0.8609, "grad_norm": 0.39816758036613464, "learning_rate": 0.0002, "epoch": 0.7212469237079573, "step": 10990}, {"loss": 0.94, "grad_norm": 0.4290051758289337, "learning_rate": 0.0002, "epoch": 0.7219031993437244, "step": 11000}, {"loss": 0.9328, "grad_norm": 0.3733605146408081, "learning_rate": 0.0002, "epoch": 0.7225594749794914, "step": 11010}, {"loss": 0.9048, "grad_norm": 0.37305396795272827, "learning_rate": 0.0002, "epoch": 0.7232157506152584, "step": 11020}, {"loss": 0.8713, "grad_norm": 0.3557985723018646, "learning_rate": 0.0002, "epoch": 0.7238720262510254, "step": 11030}, {"loss": 0.8558, "grad_norm": 0.4635949730873108, "learning_rate": 0.0002, "epoch": 0.7245283018867924, "step": 11040}, {"loss": 0.8711, "grad_norm": 0.3419910967350006, "learning_rate": 0.0002, "epoch": 0.7251845775225595, "step": 11050}, {"loss": 0.9138, "grad_norm": 0.41529953479766846, "learning_rate": 0.0002, "epoch": 0.7258408531583265, "step": 11060}, {"loss": 0.8859, "grad_norm": 0.325923353433609, "learning_rate": 0.0002, "epoch": 0.7264971287940936, "step": 11070}, {"loss": 0.9924, "grad_norm": 0.41932010650634766, "learning_rate": 0.0002, "epoch": 0.7271534044298605, "step": 11080}, {"loss": 0.9662, "grad_norm": 0.39984336495399475, "learning_rate": 0.0002, "epoch": 0.7278096800656275, "step": 11090}, {"loss": 0.8849, "grad_norm": 0.3673977851867676, "learning_rate": 0.0002, "epoch": 0.7284659557013946, "step": 11100}, {"loss": 0.9201, "grad_norm": 0.3722262680530548, "learning_rate": 0.0002, "epoch": 0.7291222313371616, "step": 11110}, {"loss": 0.9156, "grad_norm": 0.3859870135784149, "learning_rate": 0.0002, "epoch": 0.7297785069729287, "step": 11120}, {"loss": 0.899, "grad_norm": 0.3323642313480377, "learning_rate": 0.0002, "epoch": 0.7304347826086957, "step": 11130}, {"loss": 0.9244, "grad_norm": 0.37793564796447754, "learning_rate": 0.0002, "epoch": 0.7310910582444626, "step": 11140}, {"loss": 0.8767, "grad_norm": 0.44874733686447144, "learning_rate": 0.0002, "epoch": 0.7317473338802297, "step": 11150}, {"loss": 0.9078, "grad_norm": 0.4104187786579132, "learning_rate": 0.0002, "epoch": 0.7324036095159967, "step": 11160}, {"loss": 0.8828, "grad_norm": 0.3699926435947418, "learning_rate": 0.0002, "epoch": 0.7330598851517638, "step": 11170}, {"loss": 0.8909, "grad_norm": 0.41013723611831665, "learning_rate": 0.0002, "epoch": 0.7337161607875308, "step": 11180}, {"loss": 0.8934, "grad_norm": 0.370669424533844, "learning_rate": 0.0002, "epoch": 0.7343724364232977, "step": 11190}, {"loss": 0.881, "grad_norm": 0.4059436619281769, "learning_rate": 0.0002, "epoch": 0.7350287120590648, "step": 11200}, {"loss": 0.903, "grad_norm": 0.5038959383964539, "learning_rate": 0.0002, "epoch": 0.7356849876948318, "step": 11210}, {"loss": 0.9022, "grad_norm": 0.337137371301651, "learning_rate": 0.0002, "epoch": 0.7363412633305989, "step": 11220}, {"loss": 0.9171, "grad_norm": 0.412392795085907, "learning_rate": 0.0002, "epoch": 0.7369975389663659, "step": 11230}, {"loss": 0.8736, "grad_norm": 0.4415507912635803, "learning_rate": 0.0002, "epoch": 0.7376538146021329, "step": 11240}, {"loss": 0.8936, "grad_norm": 0.35013696551322937, "learning_rate": 0.0002, "epoch": 0.7383100902379, "step": 11250}, {"loss": 0.8841, "grad_norm": 0.3677300810813904, "learning_rate": 0.0002, "epoch": 0.7389663658736669, "step": 11260}, {"loss": 0.8756, "grad_norm": 0.36722511053085327, "learning_rate": 0.0002, "epoch": 0.7396226415094339, "step": 11270}, {"loss": 0.8727, "grad_norm": 0.41611534357070923, "learning_rate": 0.0002, "epoch": 0.740278917145201, "step": 11280}, {"loss": 0.921, "grad_norm": 0.4244968295097351, "learning_rate": 0.0002, "epoch": 0.740935192780968, "step": 11290}, {"loss": 0.9696, "grad_norm": 0.38986068964004517, "learning_rate": 0.0002, "epoch": 0.7415914684167351, "step": 11300}, {"loss": 0.855, "grad_norm": 0.33212459087371826, "learning_rate": 0.0002, "epoch": 0.742247744052502, "step": 11310}, {"loss": 0.9296, "grad_norm": 0.4290331304073334, "learning_rate": 0.0002, "epoch": 0.742904019688269, "step": 11320}, {"loss": 0.867, "grad_norm": 0.3750900626182556, "learning_rate": 0.0002, "epoch": 0.7435602953240361, "step": 11330}, {"loss": 0.9462, "grad_norm": 0.37758177518844604, "learning_rate": 0.0002, "epoch": 0.7442165709598031, "step": 11340}, {"loss": 0.849, "grad_norm": 0.31147271394729614, "learning_rate": 0.0002, "epoch": 0.7448728465955702, "step": 11350}, {"loss": 0.9675, "grad_norm": 0.4493428170681, "learning_rate": 0.0002, "epoch": 0.7455291222313372, "step": 11360}, {"loss": 0.9506, "grad_norm": 0.4268129765987396, "learning_rate": 0.0002, "epoch": 0.7461853978671041, "step": 11370}, {"loss": 0.9371, "grad_norm": 0.3716314733028412, "learning_rate": 0.0002, "epoch": 0.7468416735028712, "step": 11380}, {"loss": 0.8762, "grad_norm": 0.33728593587875366, "learning_rate": 0.0002, "epoch": 0.7474979491386382, "step": 11390}, {"loss": 0.8532, "grad_norm": 0.36548155546188354, "learning_rate": 0.0002, "epoch": 0.7481542247744053, "step": 11400}, {"loss": 0.8709, "grad_norm": 0.32645145058631897, "learning_rate": 0.0002, "epoch": 0.7488105004101723, "step": 11410}, {"loss": 0.9169, "grad_norm": 0.36403900384902954, "learning_rate": 0.0002, "epoch": 0.7494667760459393, "step": 11420}, {"loss": 0.9734, "grad_norm": 0.4186992049217224, "learning_rate": 0.0002, "epoch": 0.7501230516817063, "step": 11430}, {"loss": 0.8982, "grad_norm": 0.342457115650177, "learning_rate": 0.0002, "epoch": 0.7507793273174733, "step": 11440}, {"loss": 0.8785, "grad_norm": 0.3882320821285248, "learning_rate": 0.0002, "epoch": 0.7514356029532404, "step": 11450}, {"loss": 0.9492, "grad_norm": 0.4653763771057129, "learning_rate": 0.0002, "epoch": 0.7520918785890074, "step": 11460}, {"loss": 0.9357, "grad_norm": 0.3960241377353668, "learning_rate": 0.0002, "epoch": 0.7527481542247744, "step": 11470}, {"loss": 0.8912, "grad_norm": 0.38367652893066406, "learning_rate": 0.0002, "epoch": 0.7534044298605415, "step": 11480}, {"loss": 0.9367, "grad_norm": 0.39881640672683716, "learning_rate": 0.0002, "epoch": 0.7540607054963084, "step": 11490}, {"loss": 0.9018, "grad_norm": 0.40201085805892944, "learning_rate": 0.0002, "epoch": 0.7547169811320755, "step": 11500}, {"loss": 0.8668, "grad_norm": 0.38070711493492126, "learning_rate": 0.0002, "epoch": 0.7553732567678425, "step": 11510}, {"loss": 0.8542, "grad_norm": 0.334245502948761, "learning_rate": 0.0002, "epoch": 0.7560295324036095, "step": 11520}, {"loss": 0.9036, "grad_norm": 0.42725566029548645, "learning_rate": 0.0002, "epoch": 0.7566858080393766, "step": 11530}, {"loss": 0.9443, "grad_norm": 1.4140206575393677, "learning_rate": 0.0002, "epoch": 0.7573420836751436, "step": 11540}, {"loss": 0.8962, "grad_norm": 0.5802565813064575, "learning_rate": 0.0002, "epoch": 0.7579983593109105, "step": 11550}, {"loss": 0.9095, "grad_norm": 0.45892661809921265, "learning_rate": 0.0002, "epoch": 0.7586546349466776, "step": 11560}, {"loss": 0.8373, "grad_norm": 0.3837396800518036, "learning_rate": 0.0002, "epoch": 0.7593109105824446, "step": 11570}, {"loss": 0.9452, "grad_norm": 0.4158342480659485, "learning_rate": 0.0002, "epoch": 0.7599671862182117, "step": 11580}, {"loss": 0.9551, "grad_norm": 0.4374973773956299, "learning_rate": 0.0002, "epoch": 0.7606234618539787, "step": 11590}, {"loss": 0.9181, "grad_norm": 0.36342450976371765, "learning_rate": 0.0002, "epoch": 0.7612797374897456, "step": 11600}, {"loss": 0.8596, "grad_norm": 0.37019461393356323, "learning_rate": 0.0002, "epoch": 0.7619360131255127, "step": 11610}, {"loss": 0.8804, "grad_norm": 0.3745611608028412, "learning_rate": 0.0002, "epoch": 0.7625922887612797, "step": 11620}, {"loss": 0.8745, "grad_norm": 0.33488842844963074, "learning_rate": 0.0002, "epoch": 0.7632485643970468, "step": 11630}, {"loss": 0.8798, "grad_norm": 0.3700532615184784, "learning_rate": 0.0002, "epoch": 0.7639048400328138, "step": 11640}, {"loss": 0.8943, "grad_norm": 0.3722131848335266, "learning_rate": 0.0002, "epoch": 0.7645611156685808, "step": 11650}, {"loss": 0.8702, "grad_norm": 0.3463144302368164, "learning_rate": 0.0002, "epoch": 0.7652173913043478, "step": 11660}, {"loss": 0.865, "grad_norm": 0.339691162109375, "learning_rate": 0.0002, "epoch": 0.7658736669401148, "step": 11670}, {"loss": 0.9536, "grad_norm": 0.33323508501052856, "learning_rate": 0.0002, "epoch": 0.7665299425758819, "step": 11680}, {"loss": 0.8657, "grad_norm": 0.3937166631221771, "learning_rate": 0.0002, "epoch": 0.7671862182116489, "step": 11690}, {"loss": 0.8857, "grad_norm": 0.4112081527709961, "learning_rate": 0.0002, "epoch": 0.7678424938474159, "step": 11700}, {"loss": 0.9108, "grad_norm": 0.4242405593395233, "learning_rate": 0.0002, "epoch": 0.768498769483183, "step": 11710}, {"loss": 0.9148, "grad_norm": 0.33512821793556213, "learning_rate": 0.0002, "epoch": 0.7691550451189499, "step": 11720}, {"loss": 0.8599, "grad_norm": 0.5148407816886902, "learning_rate": 0.0002, "epoch": 0.769811320754717, "step": 11730}, {"loss": 0.9678, "grad_norm": 0.4562109112739563, "learning_rate": 0.0002, "epoch": 0.770467596390484, "step": 11740}, {"loss": 0.8937, "grad_norm": 0.4246354401111603, "learning_rate": 0.0002, "epoch": 0.771123872026251, "step": 11750}, {"loss": 0.8821, "grad_norm": 0.39338991045951843, "learning_rate": 0.0002, "epoch": 0.7717801476620181, "step": 11760}, {"loss": 0.8769, "grad_norm": 0.403199166059494, "learning_rate": 0.0002, "epoch": 0.7724364232977851, "step": 11770}, {"loss": 0.8914, "grad_norm": 0.4494798183441162, "learning_rate": 0.0002, "epoch": 0.7730926989335521, "step": 11780}, {"loss": 0.8633, "grad_norm": 0.3649079501628876, "learning_rate": 0.0002, "epoch": 0.7737489745693191, "step": 11790}, {"loss": 0.8684, "grad_norm": 0.3314788341522217, "learning_rate": 0.0002, "epoch": 0.7744052502050861, "step": 11800}, {"loss": 0.8756, "grad_norm": 0.36505308747291565, "learning_rate": 0.0002, "epoch": 0.7750615258408532, "step": 11810}, {"loss": 0.8778, "grad_norm": 0.45973560214042664, "learning_rate": 0.0002, "epoch": 0.7757178014766202, "step": 11820}, {"loss": 0.8845, "grad_norm": 0.3941294550895691, "learning_rate": 0.0002, "epoch": 0.7763740771123872, "step": 11830}, {"loss": 0.9033, "grad_norm": 0.34385251998901367, "learning_rate": 0.0002, "epoch": 0.7770303527481542, "step": 11840}, {"loss": 0.9595, "grad_norm": 0.35967403650283813, "learning_rate": 0.0002, "epoch": 0.7776866283839212, "step": 11850}, {"loss": 0.9193, "grad_norm": 0.4731179177761078, "learning_rate": 0.0002, "epoch": 0.7783429040196883, "step": 11860}, {"loss": 0.9344, "grad_norm": 0.38854387402534485, "learning_rate": 0.0002, "epoch": 0.7789991796554553, "step": 11870}, {"loss": 0.9056, "grad_norm": 0.3925110995769501, "learning_rate": 0.0002, "epoch": 0.7796554552912223, "step": 11880}, {"loss": 0.8818, "grad_norm": 0.35882773995399475, "learning_rate": 0.0002, "epoch": 0.7803117309269894, "step": 11890}, {"loss": 0.9097, "grad_norm": 0.4351222813129425, "learning_rate": 0.0002, "epoch": 0.7809680065627563, "step": 11900}, {"loss": 0.8866, "grad_norm": 0.39528653025627136, "learning_rate": 0.0002, "epoch": 0.7816242821985234, "step": 11910}, {"loss": 0.91, "grad_norm": 0.34876471757888794, "learning_rate": 0.0002, "epoch": 0.7822805578342904, "step": 11920}, {"loss": 0.9329, "grad_norm": 0.44766634702682495, "learning_rate": 0.0002, "epoch": 0.7829368334700574, "step": 11930}, {"loss": 0.9687, "grad_norm": 0.42268314957618713, "learning_rate": 0.0002, "epoch": 0.7835931091058245, "step": 11940}, {"loss": 0.911, "grad_norm": 0.377101868391037, "learning_rate": 0.0002, "epoch": 0.7842493847415914, "step": 11950}, {"loss": 0.8711, "grad_norm": 0.4489518404006958, "learning_rate": 0.0002, "epoch": 0.7849056603773585, "step": 11960}, {"loss": 0.9177, "grad_norm": 0.4585464596748352, "learning_rate": 0.0002, "epoch": 0.7855619360131255, "step": 11970}, {"loss": 0.901, "grad_norm": 0.41329675912857056, "learning_rate": 0.0002, "epoch": 0.7862182116488925, "step": 11980}, {"loss": 0.8993, "grad_norm": 0.42738014459609985, "learning_rate": 0.0002, "epoch": 0.7868744872846596, "step": 11990}, {"loss": 0.9206, "grad_norm": 0.40386950969696045, "learning_rate": 0.0002, "epoch": 0.7875307629204266, "step": 12000}, {"loss": 0.8702, "grad_norm": 0.3859177827835083, "learning_rate": 0.0002, "epoch": 0.7881870385561937, "step": 12010}, {"loss": 0.87, "grad_norm": 0.3537571430206299, "learning_rate": 0.0002, "epoch": 0.7888433141919606, "step": 12020}, {"loss": 0.9106, "grad_norm": 0.39319470524787903, "learning_rate": 0.0002, "epoch": 0.7894995898277276, "step": 12030}, {"loss": 0.9197, "grad_norm": 0.32762688398361206, "learning_rate": 0.0002, "epoch": 0.7901558654634947, "step": 12040}, {"loss": 0.9218, "grad_norm": 0.34617939591407776, "learning_rate": 0.0002, "epoch": 0.7908121410992617, "step": 12050}, {"loss": 0.9099, "grad_norm": 0.4012453258037567, "learning_rate": 0.0002, "epoch": 0.7914684167350287, "step": 12060}, {"loss": 0.9857, "grad_norm": 0.4265747666358948, "learning_rate": 0.0002, "epoch": 0.7921246923707957, "step": 12070}, {"loss": 0.9169, "grad_norm": 0.42164385318756104, "learning_rate": 0.0002, "epoch": 0.7927809680065627, "step": 12080}, {"loss": 0.8952, "grad_norm": 0.4054335951805115, "learning_rate": 0.0002, "epoch": 0.7934372436423298, "step": 12090}, {"loss": 0.9706, "grad_norm": 0.38486307859420776, "learning_rate": 0.0002, "epoch": 0.7940935192780968, "step": 12100}, {"loss": 0.8703, "grad_norm": 0.3842015266418457, "learning_rate": 0.0002, "epoch": 0.7947497949138638, "step": 12110}, {"loss": 0.8847, "grad_norm": 0.3772023320198059, "learning_rate": 0.0002, "epoch": 0.7954060705496309, "step": 12120}, {"loss": 0.9248, "grad_norm": 0.39477309584617615, "learning_rate": 0.0002, "epoch": 0.7960623461853978, "step": 12130}, {"loss": 0.8865, "grad_norm": 0.3458614945411682, "learning_rate": 0.0002, "epoch": 0.7967186218211649, "step": 12140}, {"loss": 0.884, "grad_norm": 0.42238670587539673, "learning_rate": 0.0002, "epoch": 0.7973748974569319, "step": 12150}, {"loss": 0.9561, "grad_norm": 0.3623220920562744, "learning_rate": 0.0002, "epoch": 0.7980311730926989, "step": 12160}, {"loss": 1.0033, "grad_norm": 0.426715224981308, "learning_rate": 0.0002, "epoch": 0.798687448728466, "step": 12170}, {"loss": 0.9158, "grad_norm": 0.3558938205242157, "learning_rate": 0.0002, "epoch": 0.799343724364233, "step": 12180}, {"loss": 0.9323, "grad_norm": 0.426761269569397, "learning_rate": 0.0002, "epoch": 0.8, "step": 12190}, {"loss": 0.9256, "grad_norm": 0.42333319783210754, "learning_rate": 0.0002, "epoch": 0.800656275635767, "step": 12200}, {"loss": 0.9307, "grad_norm": 0.42534688115119934, "learning_rate": 0.0002, "epoch": 0.801312551271534, "step": 12210}, {"loss": 0.8999, "grad_norm": 0.37565773725509644, "learning_rate": 0.0002, "epoch": 0.8019688269073011, "step": 12220}, {"loss": 0.8531, "grad_norm": 0.3591224253177643, "learning_rate": 0.0002, "epoch": 0.8026251025430681, "step": 12230}, {"loss": 0.8817, "grad_norm": 0.35887411236763, "learning_rate": 0.0002, "epoch": 0.8032813781788352, "step": 12240}, {"loss": 0.8562, "grad_norm": 0.3895672559738159, "learning_rate": 0.0002, "epoch": 0.8039376538146021, "step": 12250}, {"loss": 0.8885, "grad_norm": 0.3483835756778717, "learning_rate": 0.0002, "epoch": 0.8045939294503691, "step": 12260}, {"loss": 0.9343, "grad_norm": 0.37694090604782104, "learning_rate": 0.0002, "epoch": 0.8052502050861362, "step": 12270}, {"loss": 0.9315, "grad_norm": 0.4011424779891968, "learning_rate": 0.0002, "epoch": 0.8059064807219032, "step": 12280}, {"loss": 0.9341, "grad_norm": 0.3254278004169464, "learning_rate": 0.0002, "epoch": 0.8065627563576703, "step": 12290}, {"loss": 0.9025, "grad_norm": 0.3868531286716461, "learning_rate": 0.0002, "epoch": 0.8072190319934373, "step": 12300}, {"loss": 0.8959, "grad_norm": 0.44830775260925293, "learning_rate": 0.0002, "epoch": 0.8078753076292042, "step": 12310}, {"loss": 0.9639, "grad_norm": 0.38808006048202515, "learning_rate": 0.0002, "epoch": 0.8085315832649713, "step": 12320}, {"loss": 0.9119, "grad_norm": 0.42684856057167053, "learning_rate": 0.0002, "epoch": 0.8091878589007383, "step": 12330}, {"loss": 0.9231, "grad_norm": 0.310099333524704, "learning_rate": 0.0002, "epoch": 0.8098441345365053, "step": 12340}, {"loss": 0.8607, "grad_norm": 0.5252287983894348, "learning_rate": 0.0002, "epoch": 0.8105004101722724, "step": 12350}, {"loss": 0.8559, "grad_norm": 0.41571253538131714, "learning_rate": 0.0002, "epoch": 0.8111566858080393, "step": 12360}, {"loss": 0.8626, "grad_norm": 0.4471582770347595, "learning_rate": 0.0002, "epoch": 0.8118129614438064, "step": 12370}, {"loss": 0.923, "grad_norm": 0.3977353274822235, "learning_rate": 0.0002, "epoch": 0.8124692370795734, "step": 12380}, {"loss": 0.8792, "grad_norm": 0.3274862766265869, "learning_rate": 0.0002, "epoch": 0.8131255127153404, "step": 12390}, {"loss": 0.907, "grad_norm": 0.4114132821559906, "learning_rate": 0.0002, "epoch": 0.8137817883511075, "step": 12400}, {"loss": 0.9324, "grad_norm": 0.40929168462753296, "learning_rate": 0.0002, "epoch": 0.8144380639868745, "step": 12410}, {"loss": 0.8745, "grad_norm": 0.32346615195274353, "learning_rate": 0.0002, "epoch": 0.8150943396226416, "step": 12420}, {"loss": 0.8766, "grad_norm": 0.36344656348228455, "learning_rate": 0.0002, "epoch": 0.8157506152584085, "step": 12430}, {"loss": 0.872, "grad_norm": 0.3749464452266693, "learning_rate": 0.0002, "epoch": 0.8164068908941755, "step": 12440}, {"loss": 0.9128, "grad_norm": 0.4742373526096344, "learning_rate": 0.0002, "epoch": 0.8170631665299426, "step": 12450}, {"loss": 0.8972, "grad_norm": 0.3863218128681183, "learning_rate": 0.0002, "epoch": 0.8177194421657096, "step": 12460}, {"loss": 0.9148, "grad_norm": 0.47756487131118774, "learning_rate": 0.0002, "epoch": 0.8183757178014767, "step": 12470}, {"loss": 0.8817, "grad_norm": 0.4045886993408203, "learning_rate": 0.0002, "epoch": 0.8190319934372436, "step": 12480}, {"loss": 0.866, "grad_norm": 0.4119892418384552, "learning_rate": 0.0002, "epoch": 0.8196882690730106, "step": 12490}, {"loss": 0.866, "grad_norm": 0.3714514672756195, "learning_rate": 0.0002, "epoch": 0.8203445447087777, "step": 12500}, {"loss": 0.8943, "grad_norm": 0.38410791754722595, "learning_rate": 0.0002, "epoch": 0.8210008203445447, "step": 12510}, {"loss": 0.8857, "grad_norm": 0.3178478479385376, "learning_rate": 0.0002, "epoch": 0.8216570959803118, "step": 12520}, {"loss": 0.9356, "grad_norm": 0.4347972869873047, "learning_rate": 0.0002, "epoch": 0.8223133716160788, "step": 12530}, {"loss": 0.8926, "grad_norm": 0.4276008903980255, "learning_rate": 0.0002, "epoch": 0.8229696472518457, "step": 12540}, {"loss": 0.9213, "grad_norm": 0.36309465765953064, "learning_rate": 0.0002, "epoch": 0.8236259228876128, "step": 12550}, {"loss": 0.8847, "grad_norm": 0.45721492171287537, "learning_rate": 0.0002, "epoch": 0.8242821985233798, "step": 12560}, {"loss": 0.8655, "grad_norm": 0.37675052881240845, "learning_rate": 0.0002, "epoch": 0.8249384741591469, "step": 12570}, {"loss": 0.8717, "grad_norm": 0.41907957196235657, "learning_rate": 0.0002, "epoch": 0.8255947497949139, "step": 12580}, {"loss": 0.9465, "grad_norm": 0.36430326104164124, "learning_rate": 0.0002, "epoch": 0.8262510254306809, "step": 12590}, {"loss": 0.9244, "grad_norm": 0.45015767216682434, "learning_rate": 0.0002, "epoch": 0.8269073010664479, "step": 12600}, {"loss": 0.9076, "grad_norm": 0.369115948677063, "learning_rate": 0.0002, "epoch": 0.8275635767022149, "step": 12610}, {"loss": 0.92, "grad_norm": 0.4821915030479431, "learning_rate": 0.0002, "epoch": 0.8282198523379819, "step": 12620}, {"loss": 0.8813, "grad_norm": 0.4291541874408722, "learning_rate": 0.0002, "epoch": 0.828876127973749, "step": 12630}, {"loss": 0.9196, "grad_norm": 0.3816904127597809, "learning_rate": 0.0002, "epoch": 0.829532403609516, "step": 12640}, {"loss": 0.8821, "grad_norm": 0.3915407657623291, "learning_rate": 0.0002, "epoch": 0.8301886792452831, "step": 12650}, {"loss": 0.907, "grad_norm": 0.35964086651802063, "learning_rate": 0.0002, "epoch": 0.83084495488105, "step": 12660}, {"loss": 0.8512, "grad_norm": 0.3201609253883362, "learning_rate": 0.0002, "epoch": 0.831501230516817, "step": 12670}, {"loss": 0.9085, "grad_norm": 0.44153767824172974, "learning_rate": 0.0002, "epoch": 0.8321575061525841, "step": 12680}, {"loss": 0.859, "grad_norm": 0.37000054121017456, "learning_rate": 0.0002, "epoch": 0.8328137817883511, "step": 12690}, {"loss": 0.9413, "grad_norm": 0.39843010902404785, "learning_rate": 0.0002, "epoch": 0.8334700574241182, "step": 12700}, {"loss": 0.8787, "grad_norm": 0.3594053387641907, "learning_rate": 0.0002, "epoch": 0.8341263330598851, "step": 12710}, {"loss": 0.8542, "grad_norm": 0.38574180006980896, "learning_rate": 0.0002, "epoch": 0.8347826086956521, "step": 12720}, {"loss": 0.8746, "grad_norm": 0.4101716876029968, "learning_rate": 0.0002, "epoch": 0.8354388843314192, "step": 12730}, {"loss": 0.8442, "grad_norm": 0.3665215075016022, "learning_rate": 0.0002, "epoch": 0.8360951599671862, "step": 12740}, {"loss": 0.8898, "grad_norm": 0.39471596479415894, "learning_rate": 0.0002, "epoch": 0.8367514356029533, "step": 12750}, {"loss": 0.9026, "grad_norm": 0.35624340176582336, "learning_rate": 0.0002, "epoch": 0.8374077112387203, "step": 12760}, {"loss": 0.917, "grad_norm": 0.3838249742984772, "learning_rate": 0.0002, "epoch": 0.8380639868744872, "step": 12770}, {"loss": 0.874, "grad_norm": 0.408368319272995, "learning_rate": 0.0002, "epoch": 0.8387202625102543, "step": 12780}, {"loss": 0.934, "grad_norm": 0.46758291125297546, "learning_rate": 0.0002, "epoch": 0.8393765381460213, "step": 12790}, {"loss": 0.8603, "grad_norm": 0.35787731409072876, "learning_rate": 0.0002, "epoch": 0.8400328137817884, "step": 12800}, {"loss": 0.9026, "grad_norm": 0.39618661999702454, "learning_rate": 0.0002, "epoch": 0.8406890894175554, "step": 12810}, {"loss": 0.9307, "grad_norm": 0.44962066411972046, "learning_rate": 0.0002, "epoch": 0.8413453650533224, "step": 12820}, {"loss": 0.8756, "grad_norm": 0.36435529589653015, "learning_rate": 0.0002, "epoch": 0.8420016406890894, "step": 12830}, {"loss": 0.93, "grad_norm": 0.37484753131866455, "learning_rate": 0.0002, "epoch": 0.8426579163248564, "step": 12840}, {"loss": 0.8655, "grad_norm": 0.36679843068122864, "learning_rate": 0.0002, "epoch": 0.8433141919606235, "step": 12850}, {"loss": 0.9339, "grad_norm": 0.3948156237602234, "learning_rate": 0.0002, "epoch": 0.8439704675963905, "step": 12860}, {"loss": 0.8915, "grad_norm": 0.3789501488208771, "learning_rate": 0.0002, "epoch": 0.8446267432321575, "step": 12870}, {"loss": 0.916, "grad_norm": 0.45156800746917725, "learning_rate": 0.0002, "epoch": 0.8452830188679246, "step": 12880}, {"loss": 0.8762, "grad_norm": 0.37762370705604553, "learning_rate": 0.0002, "epoch": 0.8459392945036915, "step": 12890}, {"loss": 0.8922, "grad_norm": 0.41654065251350403, "learning_rate": 0.0002, "epoch": 0.8465955701394585, "step": 12900}, {"loss": 0.9136, "grad_norm": 0.35026174783706665, "learning_rate": 0.0002, "epoch": 0.8472518457752256, "step": 12910}, {"loss": 0.8443, "grad_norm": 0.5227314829826355, "learning_rate": 0.0002, "epoch": 0.8479081214109926, "step": 12920}, {"loss": 0.8533, "grad_norm": 0.365063339471817, "learning_rate": 0.0002, "epoch": 0.8485643970467597, "step": 12930}, {"loss": 0.9012, "grad_norm": 0.40983277559280396, "learning_rate": 0.0002, "epoch": 0.8492206726825267, "step": 12940}, {"loss": 0.8788, "grad_norm": 0.36414071917533875, "learning_rate": 0.0002, "epoch": 0.8498769483182936, "step": 12950}, {"loss": 0.8981, "grad_norm": 0.3703882694244385, "learning_rate": 0.0002, "epoch": 0.8505332239540607, "step": 12960}, {"loss": 0.9256, "grad_norm": 0.3218643367290497, "learning_rate": 0.0002, "epoch": 0.8511894995898277, "step": 12970}, {"loss": 0.8525, "grad_norm": 0.38008660078048706, "learning_rate": 0.0002, "epoch": 0.8518457752255948, "step": 12980}, {"loss": 0.8651, "grad_norm": 0.4357127547264099, "learning_rate": 0.0002, "epoch": 0.8525020508613618, "step": 12990}, {"loss": 0.8592, "grad_norm": 0.36831775307655334, "learning_rate": 0.0002, "epoch": 0.8531583264971287, "step": 13000}, {"loss": 0.8772, "grad_norm": 0.4546806812286377, "learning_rate": 0.0002, "epoch": 0.8538146021328958, "step": 13010}, {"loss": 0.865, "grad_norm": 0.3474937379360199, "learning_rate": 0.0002, "epoch": 0.8544708777686628, "step": 13020}, {"loss": 0.9098, "grad_norm": 0.438834547996521, "learning_rate": 0.0002, "epoch": 0.8551271534044299, "step": 13030}, {"loss": 0.9573, "grad_norm": 0.3401171565055847, "learning_rate": 0.0002, "epoch": 0.8557834290401969, "step": 13040}, {"loss": 0.8286, "grad_norm": 0.3887326419353485, "learning_rate": 0.0002, "epoch": 0.8564397046759639, "step": 13050}, {"loss": 0.8695, "grad_norm": 0.3489287197589874, "learning_rate": 0.0002, "epoch": 0.857095980311731, "step": 13060}, {"loss": 0.9118, "grad_norm": 0.481189489364624, "learning_rate": 0.0002, "epoch": 0.8577522559474979, "step": 13070}, {"loss": 0.9116, "grad_norm": 0.4638312757015228, "learning_rate": 0.0002, "epoch": 0.858408531583265, "step": 13080}, {"loss": 0.8775, "grad_norm": 0.43477529287338257, "learning_rate": 0.0002, "epoch": 0.859064807219032, "step": 13090}, {"loss": 0.9232, "grad_norm": 0.43358466029167175, "learning_rate": 0.0002, "epoch": 0.859721082854799, "step": 13100}, {"loss": 0.8771, "grad_norm": 0.385527104139328, "learning_rate": 0.0002, "epoch": 0.8603773584905661, "step": 13110}, {"loss": 0.8708, "grad_norm": 0.37878429889678955, "learning_rate": 0.0002, "epoch": 0.861033634126333, "step": 13120}, {"loss": 0.83, "grad_norm": 0.409476637840271, "learning_rate": 0.0002, "epoch": 0.8616899097621001, "step": 13130}, {"loss": 0.9149, "grad_norm": 0.3888716697692871, "learning_rate": 0.0002, "epoch": 0.8623461853978671, "step": 13140}, {"loss": 0.8261, "grad_norm": 0.38093528151512146, "learning_rate": 0.0002, "epoch": 0.8630024610336341, "step": 13150}, {"loss": 0.9163, "grad_norm": 0.45613282918930054, "learning_rate": 0.0002, "epoch": 0.8636587366694012, "step": 13160}, {"loss": 0.8967, "grad_norm": 0.41665518283843994, "learning_rate": 0.0002, "epoch": 0.8643150123051682, "step": 13170}, {"loss": 0.8692, "grad_norm": 0.4322538673877716, "learning_rate": 0.0002, "epoch": 0.8649712879409351, "step": 13180}, {"loss": 0.8429, "grad_norm": 0.3452875018119812, "learning_rate": 0.0002, "epoch": 0.8656275635767022, "step": 13190}, {"loss": 0.8872, "grad_norm": 0.36628788709640503, "learning_rate": 0.0002, "epoch": 0.8662838392124692, "step": 13200}, {"loss": 0.9141, "grad_norm": 0.4329487383365631, "learning_rate": 0.0002, "epoch": 0.8669401148482363, "step": 13210}, {"loss": 0.8586, "grad_norm": 0.36800137162208557, "learning_rate": 0.0002, "epoch": 0.8675963904840033, "step": 13220}, {"loss": 0.9133, "grad_norm": 0.5235224366188049, "learning_rate": 0.0002, "epoch": 0.8682526661197703, "step": 13230}, {"loss": 0.8826, "grad_norm": 0.40818873047828674, "learning_rate": 0.0002, "epoch": 0.8689089417555373, "step": 13240}, {"loss": 0.8869, "grad_norm": 0.46719685196876526, "learning_rate": 0.0002, "epoch": 0.8695652173913043, "step": 13250}, {"loss": 0.9426, "grad_norm": 0.34422767162323, "learning_rate": 0.0002, "epoch": 0.8702214930270714, "step": 13260}, {"loss": 0.8921, "grad_norm": 0.37454837560653687, "learning_rate": 0.0002, "epoch": 0.8708777686628384, "step": 13270}, {"loss": 0.89, "grad_norm": 0.39750349521636963, "learning_rate": 0.0002, "epoch": 0.8715340442986054, "step": 13280}, {"loss": 0.9256, "grad_norm": 0.3864808678627014, "learning_rate": 0.0002, "epoch": 0.8721903199343725, "step": 13290}, {"loss": 0.8663, "grad_norm": 0.36088764667510986, "learning_rate": 0.0002, "epoch": 0.8728465955701394, "step": 13300}, {"loss": 0.8334, "grad_norm": 0.384287029504776, "learning_rate": 0.0002, "epoch": 0.8735028712059065, "step": 13310}, {"loss": 0.8697, "grad_norm": 0.3988962173461914, "learning_rate": 0.0002, "epoch": 0.8741591468416735, "step": 13320}, {"loss": 0.8924, "grad_norm": 0.40126702189445496, "learning_rate": 0.0002, "epoch": 0.8748154224774405, "step": 13330}, {"loss": 0.8766, "grad_norm": 0.3931732475757599, "learning_rate": 0.0002, "epoch": 0.8754716981132076, "step": 13340}, {"loss": 0.9249, "grad_norm": 0.35348305106163025, "learning_rate": 0.0002, "epoch": 0.8761279737489746, "step": 13350}, {"loss": 0.9412, "grad_norm": 0.3603714108467102, "learning_rate": 0.0002, "epoch": 0.8767842493847416, "step": 13360}, {"loss": 0.8432, "grad_norm": 0.3853464126586914, "learning_rate": 0.0002, "epoch": 0.8774405250205086, "step": 13370}, {"loss": 0.9035, "grad_norm": 0.44406014680862427, "learning_rate": 0.0002, "epoch": 0.8780968006562756, "step": 13380}, {"loss": 0.8903, "grad_norm": 0.3563307225704193, "learning_rate": 0.0002, "epoch": 0.8787530762920427, "step": 13390}, {"loss": 0.9004, "grad_norm": 0.4507097005844116, "learning_rate": 0.0002, "epoch": 0.8794093519278097, "step": 13400}, {"loss": 0.9792, "grad_norm": 0.4069702923297882, "learning_rate": 0.0002, "epoch": 0.8800656275635766, "step": 13410}, {"loss": 0.8857, "grad_norm": 0.5091678500175476, "learning_rate": 0.0002, "epoch": 0.8807219031993437, "step": 13420}, {"loss": 0.9499, "grad_norm": 0.3784930408000946, "learning_rate": 0.0002, "epoch": 0.8813781788351107, "step": 13430}, {"loss": 0.9475, "grad_norm": 0.3937450051307678, "learning_rate": 0.0002, "epoch": 0.8820344544708778, "step": 13440}, {"loss": 0.9411, "grad_norm": 0.39602798223495483, "learning_rate": 0.0002, "epoch": 0.8826907301066448, "step": 13450}, {"loss": 0.947, "grad_norm": 0.42562225461006165, "learning_rate": 0.0002, "epoch": 0.8833470057424118, "step": 13460}, {"loss": 0.9748, "grad_norm": 0.3933939039707184, "learning_rate": 0.0002, "epoch": 0.8840032813781789, "step": 13470}, {"loss": 0.9206, "grad_norm": 0.4387489855289459, "learning_rate": 0.0002, "epoch": 0.8846595570139458, "step": 13480}, {"loss": 0.8417, "grad_norm": 0.3655209243297577, "learning_rate": 0.0002, "epoch": 0.8853158326497129, "step": 13490}, {"loss": 0.8601, "grad_norm": 0.40750762820243835, "learning_rate": 0.0002, "epoch": 0.8859721082854799, "step": 13500}, {"loss": 0.9343, "grad_norm": 0.5716604590415955, "learning_rate": 0.0002, "epoch": 0.8866283839212469, "step": 13510}, {"loss": 0.8634, "grad_norm": 0.3286498785018921, "learning_rate": 0.0002, "epoch": 0.887284659557014, "step": 13520}, {"loss": 0.8875, "grad_norm": 0.4093165993690491, "learning_rate": 0.0002, "epoch": 0.887940935192781, "step": 13530}, {"loss": 0.9056, "grad_norm": 0.37128645181655884, "learning_rate": 0.0002, "epoch": 0.888597210828548, "step": 13540}, {"loss": 0.9356, "grad_norm": 0.9630060195922852, "learning_rate": 0.0002, "epoch": 0.889253486464315, "step": 13550}, {"loss": 0.9598, "grad_norm": 0.4119563698768616, "learning_rate": 0.0002, "epoch": 0.889909762100082, "step": 13560}, {"loss": 0.8688, "grad_norm": 0.42105695605278015, "learning_rate": 0.0002, "epoch": 0.8905660377358491, "step": 13570}, {"loss": 0.9091, "grad_norm": 0.35517653822898865, "learning_rate": 0.0002, "epoch": 0.8912223133716161, "step": 13580}, {"loss": 0.9975, "grad_norm": 0.4380730986595154, "learning_rate": 0.0002, "epoch": 0.8918785890073831, "step": 13590}, {"loss": 0.8874, "grad_norm": 0.3875235915184021, "learning_rate": 0.0002, "epoch": 0.8925348646431501, "step": 13600}, {"loss": 0.8813, "grad_norm": 0.3194465935230255, "learning_rate": 0.0002, "epoch": 0.8931911402789171, "step": 13610}, {"loss": 0.8671, "grad_norm": 0.4448561370372772, "learning_rate": 0.0002, "epoch": 0.8938474159146842, "step": 13620}, {"loss": 0.88, "grad_norm": 0.4257558286190033, "learning_rate": 0.0002, "epoch": 0.8945036915504512, "step": 13630}, {"loss": 0.8844, "grad_norm": 0.3831070065498352, "learning_rate": 0.0002, "epoch": 0.8951599671862183, "step": 13640}, {"loss": 0.8836, "grad_norm": 0.3765697479248047, "learning_rate": 0.0002, "epoch": 0.8958162428219852, "step": 13650}, {"loss": 0.9015, "grad_norm": 0.3621887266635895, "learning_rate": 0.0002, "epoch": 0.8964725184577522, "step": 13660}, {"loss": 0.9147, "grad_norm": 0.4219911992549896, "learning_rate": 0.0002, "epoch": 0.8971287940935193, "step": 13670}, {"loss": 0.878, "grad_norm": 0.550961971282959, "learning_rate": 0.0002, "epoch": 0.8977850697292863, "step": 13680}, {"loss": 0.8807, "grad_norm": 0.41570132970809937, "learning_rate": 0.0002, "epoch": 0.8984413453650533, "step": 13690}, {"loss": 0.953, "grad_norm": 0.34862181544303894, "learning_rate": 0.0002, "epoch": 0.8990976210008204, "step": 13700}, {"loss": 0.8738, "grad_norm": 0.3568742871284485, "learning_rate": 0.0002, "epoch": 0.8997538966365873, "step": 13710}, {"loss": 0.9346, "grad_norm": 0.3789973258972168, "learning_rate": 0.0002, "epoch": 0.9004101722723544, "step": 13720}, {"loss": 0.8225, "grad_norm": 0.3775809109210968, "learning_rate": 0.0002, "epoch": 0.9010664479081214, "step": 13730}, {"loss": 0.9238, "grad_norm": 0.33509477972984314, "learning_rate": 0.0002, "epoch": 0.9017227235438884, "step": 13740}, {"loss": 0.915, "grad_norm": 0.4410351514816284, "learning_rate": 0.0002, "epoch": 0.9023789991796555, "step": 13750}, {"loss": 0.9066, "grad_norm": 0.44638893008232117, "learning_rate": 0.0002, "epoch": 0.9030352748154225, "step": 13760}, {"loss": 0.883, "grad_norm": 0.38487187027931213, "learning_rate": 0.0002, "epoch": 0.9036915504511895, "step": 13770}, {"loss": 0.8886, "grad_norm": 0.385796457529068, "learning_rate": 0.0002, "epoch": 0.9043478260869565, "step": 13780}, {"loss": 0.8781, "grad_norm": 0.4261656403541565, "learning_rate": 0.0002, "epoch": 0.9050041017227235, "step": 13790}, {"loss": 0.9237, "grad_norm": 0.38487741351127625, "learning_rate": 0.0002, "epoch": 0.9056603773584906, "step": 13800}, {"loss": 0.9357, "grad_norm": 0.40660005807876587, "learning_rate": 0.0002, "epoch": 0.9063166529942576, "step": 13810}, {"loss": 0.8943, "grad_norm": 0.4127330780029297, "learning_rate": 0.0002, "epoch": 0.9069729286300247, "step": 13820}, {"loss": 0.8981, "grad_norm": 0.4300757944583893, "learning_rate": 0.0002, "epoch": 0.9076292042657916, "step": 13830}, {"loss": 0.8956, "grad_norm": 0.3994467854499817, "learning_rate": 0.0002, "epoch": 0.9082854799015586, "step": 13840}, {"loss": 0.8949, "grad_norm": 0.4332261383533478, "learning_rate": 0.0002, "epoch": 0.9089417555373257, "step": 13850}, {"loss": 0.8897, "grad_norm": 0.3849696218967438, "learning_rate": 0.0002, "epoch": 0.9095980311730927, "step": 13860}, {"loss": 0.8897, "grad_norm": 0.39346274733543396, "learning_rate": 0.0002, "epoch": 0.9102543068088598, "step": 13870}, {"loss": 0.8847, "grad_norm": 0.39518049359321594, "learning_rate": 0.0002, "epoch": 0.9109105824446267, "step": 13880}, {"loss": 0.8982, "grad_norm": 0.4449180066585541, "learning_rate": 0.0002, "epoch": 0.9115668580803937, "step": 13890}, {"loss": 0.9057, "grad_norm": 0.41052138805389404, "learning_rate": 0.0002, "epoch": 0.9122231337161608, "step": 13900}, {"loss": 0.905, "grad_norm": 0.36827564239501953, "learning_rate": 0.0002, "epoch": 0.9128794093519278, "step": 13910}, {"loss": 0.9484, "grad_norm": 0.3875851631164551, "learning_rate": 0.0002, "epoch": 0.9135356849876949, "step": 13920}, {"loss": 0.9125, "grad_norm": 0.402854859828949, "learning_rate": 0.0002, "epoch": 0.9141919606234619, "step": 13930}, {"loss": 0.9014, "grad_norm": 0.3584592938423157, "learning_rate": 0.0002, "epoch": 0.9148482362592288, "step": 13940}, {"loss": 0.8432, "grad_norm": 0.3516979515552521, "learning_rate": 0.0002, "epoch": 0.9155045118949959, "step": 13950}, {"loss": 0.8747, "grad_norm": 0.4411509335041046, "learning_rate": 0.0002, "epoch": 0.9161607875307629, "step": 13960}, {"loss": 0.8758, "grad_norm": 0.47830596566200256, "learning_rate": 0.0002, "epoch": 0.9168170631665299, "step": 13970}, {"loss": 0.915, "grad_norm": 0.3669400215148926, "learning_rate": 0.0002, "epoch": 0.917473338802297, "step": 13980}, {"loss": 0.9265, "grad_norm": 0.4361213147640228, "learning_rate": 0.0002, "epoch": 0.918129614438064, "step": 13990}, {"loss": 0.9106, "grad_norm": 0.4224131405353546, "learning_rate": 0.0002, "epoch": 0.918785890073831, "step": 14000}, {"loss": 0.9138, "grad_norm": 0.5785587430000305, "learning_rate": 0.0002, "epoch": 0.919442165709598, "step": 14010}, {"loss": 0.9078, "grad_norm": 0.40788379311561584, "learning_rate": 0.0002, "epoch": 0.920098441345365, "step": 14020}, {"loss": 0.9241, "grad_norm": 0.40879732370376587, "learning_rate": 0.0002, "epoch": 0.9207547169811321, "step": 14030}, {"loss": 0.8565, "grad_norm": 0.4031982421875, "learning_rate": 0.0002, "epoch": 0.9214109926168991, "step": 14040}, {"loss": 0.8224, "grad_norm": 0.3457014560699463, "learning_rate": 0.0002, "epoch": 0.9220672682526662, "step": 14050}, {"loss": 0.89, "grad_norm": 0.38608697056770325, "learning_rate": 0.0002, "epoch": 0.9227235438884331, "step": 14060}, {"loss": 0.8637, "grad_norm": 0.39772507548332214, "learning_rate": 0.0002, "epoch": 0.9233798195242001, "step": 14070}, {"loss": 0.9102, "grad_norm": 0.3684687316417694, "learning_rate": 0.0002, "epoch": 0.9240360951599672, "step": 14080}, {"loss": 0.886, "grad_norm": 0.4479428827762604, "learning_rate": 0.0002, "epoch": 0.9246923707957342, "step": 14090}, {"loss": 0.857, "grad_norm": 0.39918506145477295, "learning_rate": 0.0002, "epoch": 0.9253486464315013, "step": 14100}, {"loss": 0.8765, "grad_norm": 0.4163782000541687, "learning_rate": 0.0002, "epoch": 0.9260049220672683, "step": 14110}, {"loss": 0.9022, "grad_norm": 0.40232136845588684, "learning_rate": 0.0002, "epoch": 0.9266611977030352, "step": 14120}, {"loss": 0.8815, "grad_norm": 0.36188459396362305, "learning_rate": 0.0002, "epoch": 0.9273174733388023, "step": 14130}, {"loss": 0.9072, "grad_norm": 0.42056623101234436, "learning_rate": 0.0002, "epoch": 0.9279737489745693, "step": 14140}, {"loss": 0.8678, "grad_norm": 0.3129708468914032, "learning_rate": 0.0002, "epoch": 0.9286300246103364, "step": 14150}, {"loss": 0.8895, "grad_norm": 0.4068623185157776, "learning_rate": 0.0002, "epoch": 0.9292863002461034, "step": 14160}, {"loss": 0.8939, "grad_norm": 0.43788865208625793, "learning_rate": 0.0002, "epoch": 0.9299425758818703, "step": 14170}, {"loss": 0.8954, "grad_norm": 0.35850921273231506, "learning_rate": 0.0002, "epoch": 0.9305988515176374, "step": 14180}, {"loss": 0.9001, "grad_norm": 0.41121411323547363, "learning_rate": 0.0002, "epoch": 0.9312551271534044, "step": 14190}, {"loss": 0.866, "grad_norm": 0.4249218702316284, "learning_rate": 0.0002, "epoch": 0.9319114027891715, "step": 14200}, {"loss": 0.9195, "grad_norm": 0.3779831826686859, "learning_rate": 0.0002, "epoch": 0.9325676784249385, "step": 14210}, {"loss": 0.8436, "grad_norm": 0.36781951785087585, "learning_rate": 0.0002, "epoch": 0.9332239540607055, "step": 14220}, {"loss": 0.8976, "grad_norm": 0.36741000413894653, "learning_rate": 0.0002, "epoch": 0.9338802296964726, "step": 14230}, {"loss": 0.9137, "grad_norm": 0.43414175510406494, "learning_rate": 0.0002, "epoch": 0.9345365053322395, "step": 14240}, {"loss": 0.8648, "grad_norm": 0.4406278431415558, "learning_rate": 0.0002, "epoch": 0.9351927809680065, "step": 14250}, {"loss": 0.8705, "grad_norm": 0.386152982711792, "learning_rate": 0.0002, "epoch": 0.9358490566037736, "step": 14260}, {"loss": 0.8697, "grad_norm": 0.3971618711948395, "learning_rate": 0.0002, "epoch": 0.9365053322395406, "step": 14270}, {"loss": 0.8925, "grad_norm": 0.3366684317588806, "learning_rate": 0.0002, "epoch": 0.9371616078753077, "step": 14280}, {"loss": 0.8702, "grad_norm": 0.42566195130348206, "learning_rate": 0.0002, "epoch": 0.9378178835110746, "step": 14290}, {"loss": 0.8767, "grad_norm": 0.3734602928161621, "learning_rate": 0.0002, "epoch": 0.9384741591468416, "step": 14300}, {"loss": 0.8889, "grad_norm": 0.4279540479183197, "learning_rate": 0.0002, "epoch": 0.9391304347826087, "step": 14310}, {"loss": 0.9319, "grad_norm": 0.43050041794776917, "learning_rate": 0.0002, "epoch": 0.9397867104183757, "step": 14320}, {"loss": 0.91, "grad_norm": 0.41909968852996826, "learning_rate": 0.0002, "epoch": 0.9404429860541428, "step": 14330}, {"loss": 0.8782, "grad_norm": 0.39651772379875183, "learning_rate": 0.0002, "epoch": 0.9410992616899098, "step": 14340}, {"loss": 0.8868, "grad_norm": 0.4171423614025116, "learning_rate": 0.0002, "epoch": 0.9417555373256767, "step": 14350}, {"loss": 0.8751, "grad_norm": 0.44906023144721985, "learning_rate": 0.0002, "epoch": 0.9424118129614438, "step": 14360}, {"loss": 0.9048, "grad_norm": 0.4213627576828003, "learning_rate": 0.0002, "epoch": 0.9430680885972108, "step": 14370}, {"loss": 0.9257, "grad_norm": 0.38457417488098145, "learning_rate": 0.0002, "epoch": 0.9437243642329779, "step": 14380}, {"loss": 0.8421, "grad_norm": 0.43104225397109985, "learning_rate": 0.0002, "epoch": 0.9443806398687449, "step": 14390}, {"loss": 0.865, "grad_norm": 0.40090736746788025, "learning_rate": 0.0002, "epoch": 0.9450369155045119, "step": 14400}, {"loss": 0.8746, "grad_norm": 0.36180031299591064, "learning_rate": 0.0002, "epoch": 0.9456931911402789, "step": 14410}, {"loss": 0.9297, "grad_norm": 0.4608926475048065, "learning_rate": 0.0002, "epoch": 0.9463494667760459, "step": 14420}, {"loss": 0.9131, "grad_norm": 0.44056418538093567, "learning_rate": 0.0002, "epoch": 0.947005742411813, "step": 14430}, {"loss": 0.8368, "grad_norm": 0.334051251411438, "learning_rate": 0.0002, "epoch": 0.94766201804758, "step": 14440}, {"loss": 0.8557, "grad_norm": 0.45580169558525085, "learning_rate": 0.0002, "epoch": 0.948318293683347, "step": 14450}, {"loss": 0.8828, "grad_norm": 0.3898446261882782, "learning_rate": 0.0002, "epoch": 0.9489745693191141, "step": 14460}, {"loss": 0.9436, "grad_norm": 0.465259850025177, "learning_rate": 0.0002, "epoch": 0.949630844954881, "step": 14470}, {"loss": 0.8907, "grad_norm": 0.3595256805419922, "learning_rate": 0.0002, "epoch": 0.9502871205906481, "step": 14480}, {"loss": 0.8593, "grad_norm": 0.3710390031337738, "learning_rate": 0.0002, "epoch": 0.9509433962264151, "step": 14490}, {"loss": 0.8373, "grad_norm": 0.4002886712551117, "learning_rate": 0.0002, "epoch": 0.9515996718621821, "step": 14500}, {"loss": 0.9091, "grad_norm": 0.4221348464488983, "learning_rate": 0.0002, "epoch": 0.9522559474979492, "step": 14510}, {"loss": 0.9205, "grad_norm": 0.4163874089717865, "learning_rate": 0.0002, "epoch": 0.9529122231337162, "step": 14520}, {"loss": 0.9125, "grad_norm": 0.40220746397972107, "learning_rate": 0.0002, "epoch": 0.9535684987694831, "step": 14530}, {"loss": 0.854, "grad_norm": 0.45836037397384644, "learning_rate": 0.0002, "epoch": 0.9542247744052502, "step": 14540}, {"loss": 0.911, "grad_norm": 0.420478492975235, "learning_rate": 0.0002, "epoch": 0.9548810500410172, "step": 14550}, {"loss": 0.8932, "grad_norm": 0.45627933740615845, "learning_rate": 0.0002, "epoch": 0.9555373256767843, "step": 14560}, {"loss": 0.8926, "grad_norm": 0.41290518641471863, "learning_rate": 0.0002, "epoch": 0.9561936013125513, "step": 14570}, {"loss": 0.8936, "grad_norm": 0.46660760045051575, "learning_rate": 0.0002, "epoch": 0.9568498769483182, "step": 14580}, {"loss": 0.9132, "grad_norm": 0.45327240228652954, "learning_rate": 0.0002, "epoch": 0.9575061525840853, "step": 14590}, {"loss": 0.8601, "grad_norm": 0.41973528265953064, "learning_rate": 0.0002, "epoch": 0.9581624282198523, "step": 14600}, {"loss": 0.8278, "grad_norm": 0.4022239148616791, "learning_rate": 0.0002, "epoch": 0.9588187038556194, "step": 14610}, {"loss": 0.8514, "grad_norm": 0.3226695954799652, "learning_rate": 0.0002, "epoch": 0.9594749794913864, "step": 14620}, {"loss": 0.8526, "grad_norm": 0.4233718812465668, "learning_rate": 0.0002, "epoch": 0.9601312551271534, "step": 14630}, {"loss": 0.9009, "grad_norm": 0.37561315298080444, "learning_rate": 0.0002, "epoch": 0.9607875307629204, "step": 14640}, {"loss": 0.8854, "grad_norm": 0.44843146204948425, "learning_rate": 0.0002, "epoch": 0.9614438063986874, "step": 14650}, {"loss": 0.8863, "grad_norm": 0.40062573552131653, "learning_rate": 0.0002, "epoch": 0.9621000820344545, "step": 14660}, {"loss": 0.8912, "grad_norm": 0.4720284938812256, "learning_rate": 0.0002, "epoch": 0.9627563576702215, "step": 14670}, {"loss": 0.89, "grad_norm": 0.43565067648887634, "learning_rate": 0.0002, "epoch": 0.9634126333059885, "step": 14680}, {"loss": 0.8612, "grad_norm": 0.41181448101997375, "learning_rate": 0.0002, "epoch": 0.9640689089417556, "step": 14690}, {"loss": 0.9464, "grad_norm": 0.4344978630542755, "learning_rate": 0.0002, "epoch": 0.9647251845775225, "step": 14700}, {"loss": 0.855, "grad_norm": 0.42558008432388306, "learning_rate": 0.0002, "epoch": 0.9653814602132896, "step": 14710}, {"loss": 0.8945, "grad_norm": 0.39161136746406555, "learning_rate": 0.0002, "epoch": 0.9660377358490566, "step": 14720}, {"loss": 0.8587, "grad_norm": 0.3784191608428955, "learning_rate": 0.0002, "epoch": 0.9666940114848236, "step": 14730}, {"loss": 0.8566, "grad_norm": 0.39039477705955505, "learning_rate": 0.0002, "epoch": 0.9673502871205907, "step": 14740}, {"loss": 0.9006, "grad_norm": 0.5566018223762512, "learning_rate": 0.0002, "epoch": 0.9680065627563577, "step": 14750}, {"loss": 0.9222, "grad_norm": 0.38877877593040466, "learning_rate": 0.0002, "epoch": 0.9686628383921247, "step": 14760}, {"loss": 0.8811, "grad_norm": 0.33369940519332886, "learning_rate": 0.0002, "epoch": 0.9693191140278917, "step": 14770}, {"loss": 0.8743, "grad_norm": 0.5109888315200806, "learning_rate": 0.0002, "epoch": 0.9699753896636587, "step": 14780}, {"loss": 0.8883, "grad_norm": 0.5364375114440918, "learning_rate": 0.0002, "epoch": 0.9706316652994258, "step": 14790}, {"loss": 0.8696, "grad_norm": 0.3702435791492462, "learning_rate": 0.0002, "epoch": 0.9712879409351928, "step": 14800}, {"loss": 0.9144, "grad_norm": 0.5094677805900574, "learning_rate": 0.0002, "epoch": 0.9719442165709598, "step": 14810}, {"loss": 0.8823, "grad_norm": 0.3565915524959564, "learning_rate": 0.0002, "epoch": 0.9726004922067268, "step": 14820}, {"loss": 0.8964, "grad_norm": 0.44756242632865906, "learning_rate": 0.0002, "epoch": 0.9732567678424938, "step": 14830}, {"loss": 0.8524, "grad_norm": 0.4272070527076721, "learning_rate": 0.0002, "epoch": 0.9739130434782609, "step": 14840}, {"loss": 0.8551, "grad_norm": 0.4303967356681824, "learning_rate": 0.0002, "epoch": 0.9745693191140279, "step": 14850}, {"loss": 0.9303, "grad_norm": 0.4528018534183502, "learning_rate": 0.0002, "epoch": 0.9752255947497949, "step": 14860}, {"loss": 0.8465, "grad_norm": 0.5745970606803894, "learning_rate": 0.0002, "epoch": 0.975881870385562, "step": 14870}, {"loss": 0.9496, "grad_norm": 0.43847736716270447, "learning_rate": 0.0002, "epoch": 0.9765381460213289, "step": 14880}, {"loss": 0.8647, "grad_norm": 0.4512104094028473, "learning_rate": 0.0002, "epoch": 0.977194421657096, "step": 14890}, {"loss": 0.9712, "grad_norm": 0.4573594331741333, "learning_rate": 0.0002, "epoch": 0.977850697292863, "step": 14900}, {"loss": 0.898, "grad_norm": 0.4297037720680237, "learning_rate": 0.0002, "epoch": 0.97850697292863, "step": 14910}, {"loss": 0.9115, "grad_norm": 0.3769957721233368, "learning_rate": 0.0002, "epoch": 0.9791632485643971, "step": 14920}, {"loss": 0.8956, "grad_norm": 0.3524457514286041, "learning_rate": 0.0002, "epoch": 0.979819524200164, "step": 14930}, {"loss": 0.9254, "grad_norm": 0.3965851664543152, "learning_rate": 0.0002, "epoch": 0.9804757998359311, "step": 14940}, {"loss": 0.8869, "grad_norm": 0.3593656122684479, "learning_rate": 0.0002, "epoch": 0.9811320754716981, "step": 14950}, {"loss": 0.8859, "grad_norm": 0.35391807556152344, "learning_rate": 0.0002, "epoch": 0.9817883511074651, "step": 14960}, {"loss": 0.9111, "grad_norm": 0.40651339292526245, "learning_rate": 0.0002, "epoch": 0.9824446267432322, "step": 14970}, {"loss": 0.8714, "grad_norm": 0.4370724558830261, "learning_rate": 0.0002, "epoch": 0.9831009023789992, "step": 14980}, {"loss": 0.8776, "grad_norm": 0.37859413027763367, "learning_rate": 0.0002, "epoch": 0.9837571780147663, "step": 14990}, {"loss": 0.8744, "grad_norm": 0.4329196512699127, "learning_rate": 0.0002, "epoch": 0.9844134536505332, "step": 15000}, {"loss": 0.8384, "grad_norm": 0.4119299054145813, "learning_rate": 0.0002, "epoch": 0.9850697292863002, "step": 15010}, {"loss": 0.8885, "grad_norm": 0.4084014892578125, "learning_rate": 0.0002, "epoch": 0.9857260049220673, "step": 15020}, {"loss": 0.9255, "grad_norm": 0.34770357608795166, "learning_rate": 0.0002, "epoch": 0.9863822805578343, "step": 15030}, {"loss": 0.9074, "grad_norm": 0.429995596408844, "learning_rate": 0.0002, "epoch": 0.9870385561936013, "step": 15040}, {"loss": 0.8251, "grad_norm": 0.4588816165924072, "learning_rate": 0.0002, "epoch": 0.9876948318293683, "step": 15050}, {"loss": 0.895, "grad_norm": 0.47414910793304443, "learning_rate": 0.0002, "epoch": 0.9883511074651353, "step": 15060}, {"loss": 0.8688, "grad_norm": 0.3370365798473358, "learning_rate": 0.0002, "epoch": 0.9890073831009024, "step": 15070}, {"loss": 0.8775, "grad_norm": 0.3697716295719147, "learning_rate": 0.0002, "epoch": 0.9896636587366694, "step": 15080}, {"loss": 0.8949, "grad_norm": 0.31965479254722595, "learning_rate": 0.0002, "epoch": 0.9903199343724364, "step": 15090}, {"loss": 0.9519, "grad_norm": 0.5081075429916382, "learning_rate": 0.0002, "epoch": 0.9909762100082035, "step": 15100}, {"loss": 0.9012, "grad_norm": 0.44397613406181335, "learning_rate": 0.0002, "epoch": 0.9916324856439704, "step": 15110}, {"loss": 0.8327, "grad_norm": 0.30696988105773926, "learning_rate": 0.0002, "epoch": 0.9922887612797375, "step": 15120}, {"loss": 0.8785, "grad_norm": 0.4071432650089264, "learning_rate": 0.0002, "epoch": 0.9929450369155045, "step": 15130}, {"loss": 0.8879, "grad_norm": 0.5356084108352661, "learning_rate": 0.0002, "epoch": 0.9936013125512715, "step": 15140}, {"loss": 0.8593, "grad_norm": 0.36654597520828247, "learning_rate": 0.0002, "epoch": 0.9942575881870386, "step": 15150}, {"loss": 0.92, "grad_norm": 0.38214483857154846, "learning_rate": 0.0002, "epoch": 0.9949138638228056, "step": 15160}, {"loss": 0.9101, "grad_norm": 0.4340892434120178, "learning_rate": 0.0002, "epoch": 0.9955701394585726, "step": 15170}, {"loss": 0.9049, "grad_norm": 0.41310828924179077, "learning_rate": 0.0002, "epoch": 0.9962264150943396, "step": 15180}, {"loss": 0.8557, "grad_norm": 0.4932044744491577, "learning_rate": 0.0002, "epoch": 0.9968826907301066, "step": 15190}, {"loss": 0.8989, "grad_norm": 0.45371273159980774, "learning_rate": 0.0002, "epoch": 0.9975389663658737, "step": 15200}, {"loss": 0.9003, "grad_norm": 0.42956778407096863, "learning_rate": 0.0002, "epoch": 0.9981952420016407, "step": 15210}, {"loss": 0.8763, "grad_norm": 0.4343477487564087, "learning_rate": 0.0002, "epoch": 0.9988515176374078, "step": 15220}, {"loss": 0.8832, "grad_norm": 0.4425382912158966, "learning_rate": 0.0002, "epoch": 0.9995077932731747, "step": 15230}]}
+{"epoch": 2.0, "step": 30475, "epoch_duration": 26246.496717453003, "total_accumulated_duration": 67246.89696502686, "gpu_info": {"GPU_0": "NVIDIA A100-PCIE-40GB"}, "memory_usage": {"avg_memory_usage": {"GPU_0": 7751.47119140625}, "peak_memory_usage": {"GPU_0": 19996.7255859375}, "avg_memory_reserved": {"GPU_0": 24388.0}, "peak_memory_reserved": {"GPU_0": 24780.0}, "total_memory": {"GPU_0": 40444.375}}, "best_checkpoint_path": "outputs-001/gemma-2-9b-it_int4_medmcqa_full_con_lr-0.0002_e-8_seq-512_lora-a-32-d-0.05-r-64_bs-1_gas-2_tf32-True_tunedata-portion-p-1.0-num-127975-sd-1/checkpoint-15237", "params": {"epochs": 8, "batch_size": 1, "learning_rate": 0.0002, "gradient_accumulation_steps": 2, "warmup_ratio": 0.03, "max_grad_norm": 0.3, "lora_alpha": 32, "lora_dropout": 0.05, "lora_r": 64, "tf32": true, "seed": 42}, "log_history": [{"loss": 2.6059, "grad_norm": 1.0711857080459595, "learning_rate": 0.0002, "epoch": 0.0006562756357670221, "step": 10}, {"loss": 1.7643, "grad_norm": 0.935492992401123, "learning_rate": 0.0002, "epoch": 0.0013125512715340443, "step": 20}, {"loss": 1.2573, "grad_norm": 0.908809244632721, "learning_rate": 0.0002, "epoch": 0.001968826907301
\ No newline at end of file