Add files using upload-large-folder tool

README.md

@@ -0,0 +1,67 @@
+---
+library_name: transformers
+model_name: out_dir_grpo_dr_geo_plus
+tags:
+- generated_from_trainer
+- trl
+- grpo
+licence: license
+---
+
+# Model Card for out_dir_grpo_dr_geo_plus
+
+This model is a fine-tuned version of [None](https://huggingface.co/None).
+It has been trained using [TRL](https://github.com/huggingface/trl).
+
+## Quick start
+
+```python
+from transformers import pipeline
+
+question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
+generator = pipeline("text-generation", model="None", device="cuda")
+output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
+print(output["generated_text"])
+```
+
+## Training procedure
+
+ 
+
+
+This model was trained with GRPO, a method introduced in [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://huggingface.co/papers/2402.03300).
+
+### Framework versions
+
+- TRL: 0.24.0.dev0
+- Transformers: 4.56.2
+- Pytorch: 2.8.0
+- Datasets: 4.0.0
+- Tokenizers: 0.22.0
+
+## Citations
+
+Cite GRPO as:
+
+```bibtex
+@article{shao2024deepseekmath,
+    title        = {{DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models}},
+    author       = {Zhihong Shao and Peiyi Wang and Qihao Zhu and Runxin Xu and Junxiao Song and Mingchuan Zhang and Y. K. Li and Y. Wu and Daya Guo},
+    year         = 2024,
+    eprint       = {arXiv:2402.03300},
+}
+
+```
+
+Cite TRL as:
+    
+```bibtex
+@misc{vonwerra2022trl,
+	title        = {{TRL: Transformer Reinforcement Learning}},
+	author       = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallou{\'e}dec},
+	year         = 2020,
+	journal      = {GitHub repository},
+	publisher    = {GitHub},
+	howpublished = {\url{https://github.com/huggingface/trl}}
+}
+```

checkpoint-1050/chat_template.jinja

@@ -0,0 +1,7 @@
+{% set image_count = namespace(value=0) %}{% set video_count = namespace(value=0) %}{% for message in messages %}{% if loop.first and message['role'] != 'system' %}<|im_start|>system
+You are a helpful assistant.<|im_end|>
+{% endif %}<|im_start|>{{ message['role'] }}
+{% if message['content'] is string %}{{ message['content'] }}<|im_end|>
+{% else %}{% for content in message['content'] %}{% if content['type'] == 'image' or 'image' in content or 'image_url' in content %}{% set image_count.value = image_count.value + 1 %}{% if add_vision_id %}Picture {{ image_count.value }}: {% endif %}<|vision_start|><|image_pad|><|vision_end|>{% elif content['type'] == 'video' or 'video' in content %}{% set video_count.value = video_count.value + 1 %}{% if add_vision_id %}Video {{ video_count.value }}: {% endif %}<|vision_start|><|video_pad|><|vision_end|>{% elif 'text' in content %}{{ content['text'] }}{% endif %}{% endfor %}<|im_end|>
+{% endif %}{% endfor %}{% if add_generation_prompt %}<|im_start|>assistant
+{% endif %}

checkpoint-1050/config.json

@@ -0,0 +1,143 @@
+{
+  "architectures": [
+    "Qwen2_5_PVLForConditionalGeneration"
+  ],
+  "attention_dropout": 0.0,
+  "dtype": "bfloat16",
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "image_token_id": 151655,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "max_position_embeddings": 128000,
+  "max_window_layers": 70,
+  "model_type": "qwen2_5_vl",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 36,
+  "num_key_value_heads": 2,
+  "pad_token_id": 151643,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": {
+    "mrope_section": [
+      16,
+      24,
+      24
+    ],
+    "rope_type": "default",
+    "type": "default"
+  },
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "text_config": {
+    "architectures": [
+      "Qwen2_5_VLForConditionalGeneration"
+    ],
+    "attention_dropout": 0.0,
+    "bos_token_id": 151643,
+    "dtype": "bfloat16",
+    "eos_token_id": 151645,
+    "hidden_act": "silu",
+    "hidden_size": 2048,
+    "image_token_id": null,
+    "initializer_range": 0.02,
+    "intermediate_size": 11008,
+    "layer_types": [
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention"
+    ],
+    "max_position_embeddings": 128000,
+    "max_window_layers": 70,
+    "model_type": "qwen2_5_vl_text",
+    "num_attention_heads": 16,
+    "num_hidden_layers": 36,
+    "num_key_value_heads": 2,
+    "rms_norm_eps": 1e-06,
+    "rope_scaling": {
+      "mrope_section": [
+        16,
+        24,
+        24
+      ],
+      "rope_type": "default",
+      "type": "default"
+    },
+    "rope_theta": 1000000.0,
+    "sliding_window": null,
+    "tie_word_embeddings": true,
+    "use_cache": true,
+    "use_sliding_window": false,
+    "video_token_id": null,
+    "vision_end_token_id": 151653,
+    "vision_start_token_id": 151652,
+    "vision_token_id": 151654,
+    "vocab_size": 151936
+  },
+  "transformers_version": "4.56.2",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "video_token_id": 151656,
+  "vision_config": {
+    "depth": 32,
+    "dtype": "bfloat16",
+    "fullatt_block_indexes": [
+      7,
+      15,
+      23,
+      31
+    ],
+    "hidden_act": "silu",
+    "hidden_size": 1280,
+    "in_channels": 3,
+    "in_chans": 3,
+    "initializer_range": 0.02,
+    "intermediate_size": 3420,
+    "model_type": "qwen2_5_vl",
+    "num_heads": 16,
+    "out_hidden_size": 2048,
+    "patch_size": 14,
+    "spatial_merge_size": 2,
+    "spatial_patch_size": 14,
+    "temporal_patch_size": 2,
+    "tokens_per_second": 2,
+    "window_size": 112
+  },
+  "vision_end_token_id": 151653,
+  "vision_start_token_id": 151652,
+  "vision_token_id": 151654,
+  "vocab_size": 151936
+}

checkpoint-1050/generation_config.json

@@ -0,0 +1,11 @@
+{
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.05,
+  "temperature": 1e-06,
+  "transformers_version": "4.56.2"
+}

checkpoint-1200/added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

checkpoint-1200/chat_template.jinja

@@ -0,0 +1,7 @@
+{% set image_count = namespace(value=0) %}{% set video_count = namespace(value=0) %}{% for message in messages %}{% if loop.first and message['role'] != 'system' %}<|im_start|>system
+You are a helpful assistant.<|im_end|>
+{% endif %}<|im_start|>{{ message['role'] }}
+{% if message['content'] is string %}{{ message['content'] }}<|im_end|>
+{% else %}{% for content in message['content'] %}{% if content['type'] == 'image' or 'image' in content or 'image_url' in content %}{% set image_count.value = image_count.value + 1 %}{% if add_vision_id %}Picture {{ image_count.value }}: {% endif %}<|vision_start|><|image_pad|><|vision_end|>{% elif content['type'] == 'video' or 'video' in content %}{% set video_count.value = video_count.value + 1 %}{% if add_vision_id %}Video {{ video_count.value }}: {% endif %}<|vision_start|><|video_pad|><|vision_end|>{% elif 'text' in content %}{{ content['text'] }}{% endif %}{% endfor %}<|im_end|>
+{% endif %}{% endfor %}{% if add_generation_prompt %}<|im_start|>assistant
+{% endif %}

checkpoint-1200/generation_config.json

@@ -0,0 +1,11 @@
+{
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.05,
+  "temperature": 1e-06,
+  "transformers_version": "4.56.2"
+}

checkpoint-1200/special_tokens_map.json

@@ -0,0 +1,31 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-1200/video_preprocessor_config.json

@@ -0,0 +1,44 @@
+{
+  "crop_size": null,
+  "data_format": "channels_first",
+  "default_to_square": true,
+  "device": null,
+  "do_center_crop": null,
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_pad": null,
+  "do_rescale": true,
+  "do_resize": true,
+  "do_sample_frames": false,
+  "fps": null,
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "input_data_format": null,
+  "max_frames": 768,
+  "max_pixels": 12845056,
+  "merge_size": 2,
+  "min_frames": 4,
+  "min_pixels": 3136,
+  "num_frames": null,
+  "patch_size": 14,
+  "processor_class": "Qwen2_5_VLProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "return_metadata": false,
+  "size": {
+    "longest_edge": 12845056,
+    "shortest_edge": 3136
+  },
+  "size_divisor": null,
+  "temporal_patch_size": 2,
+  "video_metadata": null,
+  "video_processor_type": "Qwen2VLVideoProcessor"
+}

checkpoint-800/added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

checkpoint-800/chat_template.jinja

@@ -0,0 +1,7 @@
+{% set image_count = namespace(value=0) %}{% set video_count = namespace(value=0) %}{% for message in messages %}{% if loop.first and message['role'] != 'system' %}<|im_start|>system
+You are a helpful assistant.<|im_end|>
+{% endif %}<|im_start|>{{ message['role'] }}
+{% if message['content'] is string %}{{ message['content'] }}<|im_end|>
+{% else %}{% for content in message['content'] %}{% if content['type'] == 'image' or 'image' in content or 'image_url' in content %}{% set image_count.value = image_count.value + 1 %}{% if add_vision_id %}Picture {{ image_count.value }}: {% endif %}<|vision_start|><|image_pad|><|vision_end|>{% elif content['type'] == 'video' or 'video' in content %}{% set video_count.value = video_count.value + 1 %}{% if add_vision_id %}Video {{ video_count.value }}: {% endif %}<|vision_start|><|video_pad|><|vision_end|>{% elif 'text' in content %}{{ content['text'] }}{% endif %}{% endfor %}<|im_end|>
+{% endif %}{% endfor %}{% if add_generation_prompt %}<|im_start|>assistant
+{% endif %}

checkpoint-800/generation_config.json

@@ -0,0 +1,11 @@
+{
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.05,
+  "temperature": 1e-06,
+  "transformers_version": "4.56.2"
+}

checkpoint-800/preprocessor_config.json

@@ -0,0 +1,37 @@
+{
+  "crop_size": null,
+  "data_format": "channels_first",
+  "default_to_square": true,
+  "device": null,
+  "disable_grouping": null,
+  "do_center_crop": null,
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_processor_type": "Qwen2VLImageProcessorFast",
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "input_data_format": null,
+  "max_pixels": 12845056,
+  "merge_size": 2,
+  "min_pixels": 3136,
+  "patch_size": 14,
+  "processor_class": "Qwen2_5_VLProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "return_tensors": null,
+  "size": {
+    "longest_edge": 12845056,
+    "shortest_edge": 3136
+  },
+  "temporal_patch_size": 2
+}

checkpoint-800/special_tokens_map.json

@@ -0,0 +1,31 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-800/tokenizer_config.json

@@ -0,0 +1,212 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "max_length": null,
+  "model_max_length": 131072,
+  "pad_to_multiple_of": null,
+  "pad_token": "<|endoftext|>",
+  "pad_token_type_id": 0,
+  "padding_side": "left",
+  "processor_class": "Qwen2_5_VLProcessor",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

checkpoint-800/trainer_state.json

@@ -0,0 +1,2334 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.2822865208186309,
+  "eval_steps": 500,
+  "global_step": 800,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "3d_bbox": 0.6610370799899101,
+      "adv": 0.000926484264937244,
+      "bbox": 0.8075725510716438,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.875,
+      "completions/max_terminated_length": 1592.875,
+      "completions/mean_length": 1573.1640625,
+      "completions/mean_terminated_length": 1573.1640625,
+      "completions/min_length": 1554.875,
+      "completions/min_terminated_length": 1554.875,
+      "entropy": 0.2564338077791035,
+      "epoch": 0.0028228652081863093,
+      "geo": 0.7611365988850594,
+      "grad_norm": 0.439453125,
+      "learning_rate": 7e-06,
+      "loss": -0.0002,
+      "num_tokens": 943810.0,
+      "point_2d": 0.875,
+      "point_3d": 0.638671875,
+      "reward": 0.7455703690648079,
+      "step": 8
+    },
+    {
+      "3d_bbox": 0.6985021606087685,
+      "adv": 0.0008333708265126916,
+      "bbox": 0.9951911941170692,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.25,
+      "completions/max_terminated_length": 1594.25,
+      "completions/mean_length": 1573.64453125,
+      "completions/mean_terminated_length": 1573.64453125,
+      "completions/min_length": 1552.375,
+      "completions/min_terminated_length": 1552.375,
+      "entropy": 0.2595143502112478,
+      "epoch": 0.0056457304163726185,
+      "geo": 0.8901385143399239,
+      "grad_norm": 0.341796875,
+      "learning_rate": 9.974874371859297e-06,
+      "loss": 0.0564,
+      "num_tokens": 1888015.0,
+      "point_2d": 1.0,
+      "point_3d": 0.5947265625,
+      "reward": 0.8221049904823303,
+      "step": 16
+    },
+    {
+      "3d_bbox": 0.6706115230917931,
+      "adv": -0.11667590066201683,
+      "bbox": 0.9641365110874176,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1588.75,
+      "completions/max_terminated_length": 1588.75,
+      "completions/mean_length": 1568.69921875,
+      "completions/mean_terminated_length": 1568.69921875,
+      "completions/min_length": 1549.625,
+      "completions/min_terminated_length": 1549.625,
+      "entropy": 0.2620172018650919,
+      "epoch": 0.008468595624558928,
+      "geo": 0.8719085603952408,
+      "grad_norm": 0.31640625,
+      "learning_rate": 9.934673366834172e-06,
+      "loss": 0.0434,
+      "num_tokens": 2830754.0,
+      "point_2d": 0.96484375,
+      "point_3d": 0.7060546875,
+      "reward": 0.8264116197824478,
+      "step": 24
+    },
+    {
+      "3d_bbox": 0.6768141128122807,
+      "adv": 0.00011848201484099263,
+      "bbox": 0.8640391379594803,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.875,
+      "completions/max_terminated_length": 1599.875,
+      "completions/mean_length": 1579.40625,
+      "completions/mean_terminated_length": 1579.40625,
+      "completions/min_length": 1563.0,
+      "completions/min_terminated_length": 1563.0,
+      "entropy": 0.26498199813067913,
+      "epoch": 0.011291460832745237,
+      "geo": 0.8661364912986755,
+      "grad_norm": 0.36328125,
+      "learning_rate": 9.894472361809046e-06,
+      "loss": -0.0005,
+      "num_tokens": 3776674.0,
+      "point_2d": 0.8642578125,
+      "point_3d": 0.603515625,
+      "reward": 0.752156674861908,
+      "step": 32
+    },
+    {
+      "3d_bbox": 0.6714089289307594,
+      "adv": -0.000722882674608627,
+      "bbox": 0.9624817371368408,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1593.75,
+      "completions/max_terminated_length": 1593.75,
+      "completions/mean_length": 1575.52734375,
+      "completions/mean_terminated_length": 1575.52734375,
+      "completions/min_length": 1559.875,
+      "completions/min_terminated_length": 1559.875,
+      "entropy": 0.2738626003265381,
+      "epoch": 0.014114326040931546,
+      "geo": 0.8932898789644241,
+      "grad_norm": 0.341796875,
+      "learning_rate": 9.85427135678392e-06,
+      "loss": 0.0002,
+      "num_tokens": 4721169.0,
+      "point_2d": 0.99609375,
+      "point_3d": 0.6064453125,
+      "reward": 0.8091074377298355,
+      "step": 40
+    },
+    {
+      "3d_bbox": 0.6575617864727974,
+      "adv": -0.0007460190936399158,
+      "bbox": 0.9782034233212471,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.75,
+      "completions/max_terminated_length": 1598.75,
+      "completions/mean_length": 1581.515625,
+      "completions/mean_terminated_length": 1581.515625,
+      "completions/min_length": 1563.875,
+      "completions/min_terminated_length": 1563.875,
+      "entropy": 0.26874606031924486,
+      "epoch": 0.016937191249117856,
+      "geo": 0.922828696668148,
+      "grad_norm": 0.384765625,
+      "learning_rate": 9.814070351758794e-06,
+      "loss": 0.0001,
+      "num_tokens": 5667501.0,
+      "point_2d": 0.9404296875,
+      "point_3d": 0.6171875,
+      "reward": 0.7983455955982208,
+      "step": 48
+    },
+    {
+      "3d_bbox": 0.7516707628965378,
+      "adv": 0.0002944665375252953,
+      "bbox": 0.870270783547312,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1605.5,
+      "completions/max_terminated_length": 1605.5,
+      "completions/mean_length": 1583.109375,
+      "completions/mean_terminated_length": 1583.109375,
+      "completions/min_length": 1563.125,
+      "completions/min_terminated_length": 1563.125,
+      "entropy": 0.27441642340272665,
+      "epoch": 0.019760056457304165,
+      "geo": 0.8707154765725136,
+      "grad_norm": 0.35546875,
+      "learning_rate": 9.773869346733669e-06,
+      "loss": 0.0002,
+      "num_tokens": 6614089.0,
+      "point_2d": 0.875,
+      "point_3d": 0.728515625,
+      "reward": 0.8063642866909504,
+      "step": 56
+    },
+    {
+      "3d_bbox": 0.7704295367002487,
+      "adv": 0.0006564971263287589,
+      "bbox": 0.8711516009643674,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1604.75,
+      "completions/max_terminated_length": 1604.75,
+      "completions/mean_length": 1582.09375,
+      "completions/mean_terminated_length": 1582.09375,
+      "completions/min_length": 1562.25,
+      "completions/min_terminated_length": 1562.25,
+      "entropy": 0.2721484359353781,
+      "epoch": 0.022582921665490474,
+      "geo": 0.8998556286096573,
+      "grad_norm": 0.373046875,
+      "learning_rate": 9.733668341708543e-06,
+      "loss": 0.0156,
+      "num_tokens": 7560513.0,
+      "point_2d": 0.875,
+      "point_3d": 0.7880859375,
+      "reward": 0.8261667564511299,
+      "step": 64
+    },
+    {
+      "3d_bbox": 0.6965150758624077,
+      "adv": 5.7399442994210403e-05,
+      "bbox": 0.9924400746822357,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1603.875,
+      "completions/max_terminated_length": 1603.875,
+      "completions/mean_length": 1580.8203125,
+      "completions/mean_terminated_length": 1580.8203125,
+      "completions/min_length": 1562.0,
+      "completions/min_terminated_length": 1562.0,
+      "entropy": 0.2694994225166738,
+      "epoch": 0.025405786873676783,
+      "geo": 0.7992948293685913,
+      "grad_norm": 0.40625,
+      "learning_rate": 9.693467336683417e-06,
+      "loss": -0.0004,
+      "num_tokens": 8506787.0,
+      "point_2d": 0.875,
+      "point_3d": 0.6201171875,
+      "reward": 0.796018086373806,
+      "step": 72
+    },
+    {
+      "3d_bbox": 0.9022366181015968,
+      "adv": 0.000732647989934776,
+      "bbox": 0.9670402258634567,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.5,
+      "completions/max_terminated_length": 1598.5,
+      "completions/mean_length": 1577.83203125,
+      "completions/mean_terminated_length": 1577.83203125,
+      "completions/min_length": 1560.0,
+      "completions/min_terminated_length": 1560.0,
+      "entropy": 0.2729479926638305,
+      "epoch": 0.028228652081863093,
+      "geo": 0.770210200920701,
+      "grad_norm": 0.28515625,
+      "learning_rate": 9.653266331658291e-06,
+      "loss": 0.0285,
+      "num_tokens": 9452032.0,
+      "point_2d": 0.7578125,
+      "point_3d": 0.8740234375,
+      "reward": 0.8752781972289085,
+      "step": 80
+    },
+    {
+      "3d_bbox": 0.8243749290704727,
+      "adv": 0.0004942354327113208,
+      "bbox": 0.9483200311660767,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.0,
+      "completions/max_terminated_length": 1599.0,
+      "completions/mean_length": 1578.5703125,
+      "completions/mean_terminated_length": 1578.5703125,
+      "completions/min_length": 1564.875,
+      "completions/min_terminated_length": 1564.875,
+      "entropy": 0.27196697471663356,
+      "epoch": 0.031051517290049402,
+      "geo": 0.8520620986819267,
+      "grad_norm": 0.337890625,
+      "learning_rate": 9.613065326633166e-06,
+      "loss": 0.0,
+      "num_tokens": 10397602.0,
+      "point_2d": 0.84375,
+      "point_3d": 0.7431640625,
+      "reward": 0.8399022594094276,
+      "step": 88
+    },
+    {
+      "3d_bbox": 0.7930564805865288,
+      "adv": -0.05789744130743202,
+      "bbox": 0.9452124014496803,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1600.125,
+      "completions/max_terminated_length": 1600.125,
+      "completions/mean_length": 1579.25390625,
+      "completions/mean_terminated_length": 1579.25390625,
+      "completions/min_length": 1560.25,
+      "completions/min_terminated_length": 1560.25,
+      "entropy": 0.2732193539850414,
+      "epoch": 0.03387438249823571,
+      "geo": 0.8466203399002552,
+      "grad_norm": 0.35546875,
+      "learning_rate": 9.57286432160804e-06,
+      "loss": 0.0143,
+      "num_tokens": 11343267.0,
+      "point_2d": 0.7138671875,
+      "point_3d": 0.8203125,
+      "reward": 0.8181121572852135,
+      "step": 96
+    },
+    {
+      "3d_bbox": 0.6107643470168114,
+      "adv": -0.0019174332555849105,
+      "bbox": 0.9693493247032166,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1605.375,
+      "completions/max_terminated_length": 1605.375,
+      "completions/mean_length": 1580.6875,
+      "completions/mean_terminated_length": 1580.6875,
+      "completions/min_length": 1559.75,
+      "completions/min_terminated_length": 1559.75,
+      "entropy": 0.2712484383955598,
+      "epoch": 0.03669724770642202,
+      "geo": 0.760684534907341,
+      "grad_norm": 0.392578125,
+      "learning_rate": 9.532663316582916e-06,
+      "loss": 0.0006,
+      "num_tokens": 12289259.0,
+      "point_2d": 0.849609375,
+      "point_3d": 0.7109375,
+      "reward": 0.7851651459932327,
+      "step": 104
+    },
+    {
+      "3d_bbox": 0.7011105790734291,
+      "adv": 0.000332610031364311,
+      "bbox": 0.997136726975441,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.5,
+      "completions/max_terminated_length": 1601.5,
+      "completions/mean_length": 1578.38671875,
+      "completions/mean_terminated_length": 1578.38671875,
+      "completions/min_length": 1560.0,
+      "completions/min_terminated_length": 1560.0,
+      "entropy": 0.2741631781682372,
+      "epoch": 0.03952011291460833,
+      "geo": 0.868082121014595,
+      "grad_norm": 0.349609375,
+      "learning_rate": 9.49246231155779e-06,
+      "loss": -0.0005,
+      "num_tokens": 13234494.0,
+      "point_2d": 0.875,
+      "point_3d": 0.759765625,
+      "reward": 0.8332532122731209,
+      "step": 112
+    },
+    {
+      "3d_bbox": 0.6414853371679783,
+      "adv": 6.49127487122314e-05,
+      "bbox": 0.9946233779191971,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.875,
+      "completions/max_terminated_length": 1599.875,
+      "completions/mean_length": 1580.48828125,
+      "completions/mean_terminated_length": 1580.48828125,
+      "completions/min_length": 1562.5,
+      "completions/min_terminated_length": 1562.5,
+      "entropy": 0.2765682805329561,
+      "epoch": 0.04234297812279464,
+      "geo": 0.7812720276415348,
+      "grad_norm": 0.392578125,
+      "learning_rate": 9.452261306532664e-06,
+      "loss": -0.0002,
+      "num_tokens": 14180739.0,
+      "point_2d": 0.875,
+      "point_3d": 0.7275390625,
+      "reward": 0.8096619322896004,
+      "step": 120
+    },
+    {
+      "3d_bbox": 0.696478821337223,
+      "adv": 0.00018893864194069465,
+      "bbox": 0.9915411844849586,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1600.75,
+      "completions/max_terminated_length": 1600.75,
+      "completions/mean_length": 1581.48828125,
+      "completions/mean_terminated_length": 1581.48828125,
+      "completions/min_length": 1566.25,
+      "completions/min_terminated_length": 1566.25,
+      "entropy": 0.27939224475994706,
+      "epoch": 0.04516584333098095,
+      "geo": 0.9225678443908691,
+      "grad_norm": 0.359375,
+      "learning_rate": 9.412060301507539e-06,
+      "loss": 0.0004,
+      "num_tokens": 15126808.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8037109375,
+      "reward": 0.8729327321052551,
+      "step": 128
+    },
+    {
+      "3d_bbox": 0.7811713069677353,
+      "adv": -0.00017097587488024146,
+      "bbox": 0.983598880469799,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.0,
+      "completions/max_terminated_length": 1599.0,
+      "completions/mean_length": 1575.9765625,
+      "completions/mean_terminated_length": 1575.9765625,
+      "completions/min_length": 1560.875,
+      "completions/min_terminated_length": 1560.875,
+      "entropy": 0.2712345658801496,
+      "epoch": 0.04798870853916726,
+      "geo": 0.829373586922884,
+      "grad_norm": 0.365234375,
+      "learning_rate": 9.371859296482413e-06,
+      "loss": 0.0002,
+      "num_tokens": 16071682.0,
+      "point_2d": 0.75,
+      "point_3d": 0.76171875,
+      "reward": 0.819122239947319,
+      "step": 136
+    },
+    {
+      "3d_bbox": 0.6566238105297089,
+      "adv": 5.832723485355018e-05,
+      "bbox": 0.9859664067626,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1597.5,
+      "completions/max_terminated_length": 1597.5,
+      "completions/mean_length": 1580.12890625,
+      "completions/mean_terminated_length": 1580.12890625,
+      "completions/min_length": 1565.0,
+      "completions/min_terminated_length": 1565.0,
+      "entropy": 0.2754409732297063,
+      "epoch": 0.05081157374735357,
+      "geo": 0.8895541727542877,
+      "grad_norm": 0.318359375,
+      "learning_rate": 9.331658291457287e-06,
+      "loss": 0.0218,
+      "num_tokens": 17017499.0,
+      "point_2d": 1.0,
+      "point_3d": 0.6328125,
+      "reward": 0.8188506811857224,
+      "step": 144
+    },
+    {
+      "3d_bbox": 0.655942264944315,
+      "adv": 0.00024080171715468168,
+      "bbox": 0.9830196872353554,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1600.125,
+      "completions/max_terminated_length": 1600.125,
+      "completions/mean_length": 1581.3828125,
+      "completions/mean_terminated_length": 1581.3828125,
+      "completions/min_length": 1564.0,
+      "completions/min_terminated_length": 1564.0,
+      "entropy": 0.2764031286351383,
+      "epoch": 0.053634438955539876,
+      "geo": 0.8772727139294147,
+      "grad_norm": 0.294921875,
+      "learning_rate": 9.291457286432161e-06,
+      "loss": 0.015,
+      "num_tokens": 17963757.0,
+      "point_2d": 0.875,
+      "point_3d": 0.7880859375,
+      "reward": 0.8255119547247887,
+      "step": 152
+    },
+    {
+      "3d_bbox": 0.8087940737605095,
+      "adv": -0.062093503400319605,
+      "bbox": 0.9799230694770813,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.5,
+      "completions/max_terminated_length": 1598.5,
+      "completions/mean_length": 1579.2421875,
+      "completions/mean_terminated_length": 1579.2421875,
+      "completions/min_length": 1560.25,
+      "completions/min_terminated_length": 1560.25,
+      "entropy": 0.27893486013635993,
+      "epoch": 0.056457304163726185,
+      "geo": 0.8533996716141701,
+      "grad_norm": 0.314453125,
+      "learning_rate": 9.251256281407036e-06,
+      "loss": 0.0157,
+      "num_tokens": 18909307.0,
+      "point_2d": 0.859375,
+      "point_3d": 0.814453125,
+      "reward": 0.8656363189220428,
+      "step": 160
+    },
+    {
+      "3d_bbox": 0.7417622283101082,
+      "adv": 0.00038278038937278325,
+      "bbox": 0.9947961643338203,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.875,
+      "completions/max_terminated_length": 1596.875,
+      "completions/mean_length": 1578.4765625,
+      "completions/mean_terminated_length": 1578.4765625,
+      "completions/min_length": 1562.375,
+      "completions/min_terminated_length": 1562.375,
+      "entropy": 0.2772809020243585,
+      "epoch": 0.059280169371912494,
+      "geo": 0.8446554020047188,
+      "grad_norm": 0.337890625,
+      "learning_rate": 9.21105527638191e-06,
+      "loss": -0.0001,
+      "num_tokens": 19854893.0,
+      "point_2d": 0.875,
+      "point_3d": 0.7666015625,
+      "reward": 0.844539999961853,
+      "step": 168
+    },
+    {
+      "3d_bbox": 0.7199448123574257,
+      "adv": -0.00025279237161157653,
+      "bbox": 0.9939833134412766,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.375,
+      "completions/max_terminated_length": 1599.375,
+      "completions/mean_length": 1577.875,
+      "completions/mean_terminated_length": 1577.875,
+      "completions/min_length": 1560.875,
+      "completions/min_terminated_length": 1560.875,
+      "entropy": 0.2748410953208804,
+      "epoch": 0.062103034580098804,
+      "geo": 0.8426482602953911,
+      "grad_norm": 0.353515625,
+      "learning_rate": 9.170854271356784e-06,
+      "loss": 0.0001,
+      "num_tokens": 20800093.0,
+      "point_2d": 0.8935546875,
+      "point_3d": 0.7666015625,
+      "reward": 0.8435210958123207,
+      "step": 176
+    },
+    {
+      "3d_bbox": 0.8168098330497742,
+      "adv": 0.0003567840885807527,
+      "bbox": 0.8971930183470249,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.5,
+      "completions/max_terminated_length": 1601.5,
+      "completions/mean_length": 1581.01171875,
+      "completions/mean_terminated_length": 1581.01171875,
+      "completions/min_length": 1562.625,
+      "completions/min_terminated_length": 1562.625,
+      "entropy": 0.27716846112161875,
+      "epoch": 0.06492589978828511,
+      "geo": 0.8458116948604584,
+      "grad_norm": 0.376953125,
+      "learning_rate": 9.13065326633166e-06,
+      "loss": -0.0,
+      "num_tokens": 21746232.0,
+      "point_2d": 0.736328125,
+      "point_3d": 0.859375,
+      "reward": 0.8274264857172966,
+      "step": 184
+    },
+    {
+      "3d_bbox": 0.7658436857163906,
+      "adv": -0.00043575108065851964,
+      "bbox": 0.994478665292263,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.5,
+      "completions/max_terminated_length": 1599.5,
+      "completions/mean_length": 1577.2265625,
+      "completions/mean_terminated_length": 1577.2265625,
+      "completions/min_length": 1558.25,
+      "completions/min_terminated_length": 1558.25,
+      "entropy": 0.28411462996155024,
+      "epoch": 0.06774876499647142,
+      "geo": 0.8858354687690735,
+      "grad_norm": 0.404296875,
+      "learning_rate": 9.090452261306534e-06,
+      "loss": -0.0,
+      "num_tokens": 22691162.0,
+      "point_2d": 1.0,
+      "point_3d": 0.83203125,
+      "reward": 0.8980883881449699,
+      "step": 192
+    },
+    {
+      "3d_bbox": 0.689879797399044,
+      "adv": 0.0008139211470279406,
+      "bbox": 0.8756355233490467,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.25,
+      "completions/max_terminated_length": 1595.25,
+      "completions/mean_length": 1579.328125,
+      "completions/mean_terminated_length": 1579.328125,
+      "completions/min_length": 1562.125,
+      "completions/min_terminated_length": 1562.125,
+      "entropy": 0.27792059257626534,
+      "epoch": 0.07057163020465773,
+      "geo": 0.8776071295142174,
+      "grad_norm": 0.28125,
+      "learning_rate": 9.050251256281409e-06,
+      "loss": -0.0006,
+      "num_tokens": 23637118.0,
+      "point_2d": 0.875,
+      "point_3d": 0.765625,
+      "reward": 0.8015350699424744,
+      "step": 200
+    },
+    {
+      "3d_bbox": 0.7628588750958443,
+      "adv": 0.0005178588535272866,
+      "bbox": 0.9862369671463966,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1603.5,
+      "completions/max_terminated_length": 1603.5,
+      "completions/mean_length": 1582.5,
+      "completions/mean_terminated_length": 1582.5,
+      "completions/min_length": 1561.75,
+      "completions/min_terminated_length": 1561.75,
+      "entropy": 0.2777437614277005,
+      "epoch": 0.07339449541284404,
+      "geo": 0.7896297052502632,
+      "grad_norm": 0.318359375,
+      "learning_rate": 9.010050251256283e-06,
+      "loss": -0.0003,
+      "num_tokens": 24583790.0,
+      "point_2d": 0.75,
+      "point_3d": 0.822265625,
+      "reward": 0.8303403630852699,
+      "step": 208
+    },
+    {
+      "3d_bbox": 0.6901299897581339,
+      "adv": 1.327232166659087e-05,
+      "bbox": 0.9742533341050148,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.375,
+      "completions/max_terminated_length": 1595.375,
+      "completions/mean_length": 1580.3828125,
+      "completions/mean_terminated_length": 1580.3828125,
+      "completions/min_length": 1566.125,
+      "completions/min_terminated_length": 1566.125,
+      "entropy": 0.2789337821304798,
+      "epoch": 0.07621736062103035,
+      "geo": 0.9093400537967682,
+      "grad_norm": 0.33984375,
+      "learning_rate": 8.969849246231157e-06,
+      "loss": 0.0001,
+      "num_tokens": 25529904.0,
+      "point_2d": 1.0,
+      "point_3d": 0.724609375,
+      "reward": 0.8472481593489647,
+      "step": 216
+    },
+    {
+      "3d_bbox": 0.8317120671272278,
+      "adv": 0.0007813970478309784,
+      "bbox": 0.99237409979105,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.375,
+      "completions/max_terminated_length": 1592.375,
+      "completions/mean_length": 1576.484375,
+      "completions/mean_terminated_length": 1576.484375,
+      "completions/min_length": 1559.125,
+      "completions/min_terminated_length": 1559.125,
+      "entropy": 0.281503742095083,
+      "epoch": 0.07904022582921666,
+      "geo": 0.9148535653948784,
+      "grad_norm": 0.33203125,
+      "learning_rate": 8.92964824120603e-06,
+      "loss": -0.0005,
+      "num_tokens": 26474884.0,
+      "point_2d": 1.0,
+      "point_3d": 0.859375,
+      "reward": 0.9208653047680855,
+      "step": 224
+    },
+    {
+      "3d_bbox": 0.8539343103766441,
+      "adv": 0.0006083497164581786,
+      "bbox": 0.9947174340486526,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1607.25,
+      "completions/max_terminated_length": 1607.25,
+      "completions/mean_length": 1582.21875,
+      "completions/mean_terminated_length": 1582.21875,
+      "completions/min_length": 1562.75,
+      "completions/min_terminated_length": 1562.75,
+      "entropy": 0.28075447864830494,
+      "epoch": 0.08186309103740297,
+      "geo": 0.9205377101898193,
+      "grad_norm": 0.3828125,
+      "learning_rate": 8.889447236180904e-06,
+      "loss": 0.0151,
+      "num_tokens": 27421356.0,
+      "point_2d": 0.9423828125,
+      "point_3d": 0.919921875,
+      "reward": 0.9277391210198402,
+      "step": 232
+    },
+    {
+      "3d_bbox": 0.8143627047538757,
+      "adv": 0.0002584877729532309,
+      "bbox": 0.9961168393492699,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1604.375,
+      "completions/max_terminated_length": 1604.375,
+      "completions/mean_length": 1581.84375,
+      "completions/mean_terminated_length": 1581.84375,
+      "completions/min_length": 1562.0,
+      "completions/min_terminated_length": 1562.0,
+      "entropy": 0.2808487247675657,
+      "epoch": 0.08468595624558928,
+      "geo": 0.9482192173600197,
+      "grad_norm": 0.3125,
+      "learning_rate": 8.84924623115578e-06,
+      "loss": 0.015,
+      "num_tokens": 28367588.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8505859375,
+      "reward": 0.9152663722634315,
+      "step": 240
+    },
+    {
+      "3d_bbox": 0.7562066540122032,
+      "adv": 0.0011895421193912625,
+      "bbox": 0.9870459213852882,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1605.625,
+      "completions/max_terminated_length": 1605.625,
+      "completions/mean_length": 1582.22265625,
+      "completions/mean_terminated_length": 1582.22265625,
+      "completions/min_length": 1560.125,
+      "completions/min_terminated_length": 1560.125,
+      "entropy": 0.2828774293884635,
+      "epoch": 0.08750882145377559,
+      "geo": 0.8906307443976402,
+      "grad_norm": 0.310546875,
+      "learning_rate": 8.809045226130654e-06,
+      "loss": -0.0001,
+      "num_tokens": 29314053.0,
+      "point_2d": 1.0,
+      "point_3d": 0.84765625,
+      "reward": 0.8977271988987923,
+      "step": 248
+    },
+    {
+      "3d_bbox": 0.662892609834671,
+      "adv": -0.0004179688154977157,
+      "bbox": 0.9923923686146736,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.875,
+      "completions/max_terminated_length": 1598.875,
+      "completions/mean_length": 1579.796875,
+      "completions/mean_terminated_length": 1579.796875,
+      "completions/min_length": 1563.125,
+      "completions/min_terminated_length": 1563.125,
+      "entropy": 0.27959624165669084,
+      "epoch": 0.0903316866619619,
+      "geo": 0.9055434167385101,
+      "grad_norm": 0.259765625,
+      "learning_rate": 8.768844221105528e-06,
+      "loss": -0.0002,
+      "num_tokens": 30259921.0,
+      "point_2d": 0.7509765625,
+      "point_3d": 0.8154296875,
+      "reward": 0.805422805249691,
+      "step": 256
+    },
+    {
+      "3d_bbox": 0.7727803811430931,
+      "adv": -0.0010380800276834634,
+      "bbox": 0.9807735681533813,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1604.375,
+      "completions/max_terminated_length": 1604.375,
+      "completions/mean_length": 1581.1796875,
+      "completions/mean_terminated_length": 1581.1796875,
+      "completions/min_length": 1564.375,
+      "completions/min_terminated_length": 1564.375,
+      "entropy": 0.28243178501725197,
+      "epoch": 0.0931545518701482,
+      "geo": 0.783092018100433,
+      "grad_norm": 0.3203125,
+      "learning_rate": 8.728643216080403e-06,
+      "loss": 0.0003,
+      "num_tokens": 31205887.0,
+      "point_2d": 0.75,
+      "point_3d": 0.8544921875,
+      "reward": 0.8395115286111832,
+      "step": 264
+    },
+    {
+      "3d_bbox": 0.7379852011799812,
+      "adv": -0.0005326499958755448,
+      "bbox": 0.9934581220149994,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.875,
+      "completions/max_terminated_length": 1598.875,
+      "completions/mean_length": 1579.44921875,
+      "completions/mean_terminated_length": 1579.44921875,
+      "completions/min_length": 1560.375,
+      "completions/min_terminated_length": 1560.375,
+      "entropy": 0.2767978576011956,
+      "epoch": 0.09597741707833451,
+      "geo": 0.9199443385004997,
+      "grad_norm": 0.40625,
+      "learning_rate": 8.688442211055277e-06,
+      "loss": 0.0002,
+      "num_tokens": 32151650.0,
+      "point_2d": 0.9990234375,
+      "point_3d": 0.8896484375,
+      "reward": 0.9050287976861,
+      "step": 272
+    },
+    {
+      "3d_bbox": 0.6362542062997818,
+      "adv": 0.0003683161558001302,
+      "bbox": 0.9273895137012005,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1603.375,
+      "completions/max_terminated_length": 1603.375,
+      "completions/mean_length": 1583.1875,
+      "completions/mean_terminated_length": 1583.1875,
+      "completions/min_length": 1566.25,
+      "completions/min_terminated_length": 1566.25,
+      "entropy": 0.28121166303753853,
+      "epoch": 0.09880028228652082,
+      "geo": 0.6698016822338104,
+      "grad_norm": 0.419921875,
+      "learning_rate": 8.648241206030151e-06,
+      "loss": 0.0104,
+      "num_tokens": 33098514.0,
+      "point_2d": 0.748046875,
+      "point_3d": 0.8564453125,
+      "reward": 0.7920339703559875,
+      "step": 280
+    },
+    {
+      "3d_bbox": 0.8045329824090004,
+      "adv": 0.0005392557122831931,
+      "bbox": 0.9957676902413368,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.75,
+      "completions/max_terminated_length": 1598.75,
+      "completions/mean_length": 1582.8515625,
+      "completions/mean_terminated_length": 1582.8515625,
+      "completions/min_length": 1564.75,
+      "completions/min_terminated_length": 1564.75,
+      "entropy": 0.2840573275461793,
+      "epoch": 0.10162314749470713,
+      "geo": 0.894830085337162,
+      "grad_norm": 0.263671875,
+      "learning_rate": 8.608040201005025e-06,
+      "loss": -0.0003,
+      "num_tokens": 34045156.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8974609375,
+      "reward": 0.9244403913617134,
+      "step": 288
+    },
+    {
+      "3d_bbox": 0.6487490832805634,
+      "adv": -0.036328162997961044,
+      "bbox": 0.9537674784660339,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1606.125,
+      "completions/max_terminated_length": 1606.125,
+      "completions/mean_length": 1583.43359375,
+      "completions/mean_terminated_length": 1583.43359375,
+      "completions/min_length": 1564.25,
+      "completions/min_terminated_length": 1564.25,
+      "entropy": 0.2789871282875538,
+      "epoch": 0.10444601270289344,
+      "geo": 0.8219972774386406,
+      "grad_norm": 0.3671875,
+      "learning_rate": 8.5678391959799e-06,
+      "loss": 0.0091,
+      "num_tokens": 34991979.0,
+      "point_2d": 0.875,
+      "point_3d": 0.8125,
+      "reward": 0.8225041329860687,
+      "step": 296
+    },
+    {
+      "3d_bbox": 0.7852863818407059,
+      "adv": -2.922765997936949e-05,
+      "bbox": 0.9355472847819328,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.5,
+      "completions/max_terminated_length": 1601.5,
+      "completions/mean_length": 1579.7890625,
+      "completions/mean_terminated_length": 1579.7890625,
+      "completions/min_length": 1561.5,
+      "completions/min_terminated_length": 1561.5,
+      "entropy": 0.2808959139510989,
+      "epoch": 0.10726887791107975,
+      "geo": 0.7100357487797737,
+      "grad_norm": 0.333984375,
+      "learning_rate": 8.527638190954774e-06,
+      "loss": 0.0003,
+      "num_tokens": 35937853.0,
+      "point_2d": 0.6728515625,
+      "point_3d": 0.798828125,
+      "reward": 0.7981283441185951,
+      "step": 304
+    },
+    {
+      "3d_bbox": 0.6897341758012772,
+      "adv": -0.0010775219143397408,
+      "bbox": 0.9957167953252792,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.0,
+      "completions/max_terminated_length": 1598.0,
+      "completions/mean_length": 1578.40234375,
+      "completions/mean_terminated_length": 1578.40234375,
+      "completions/min_length": 1560.625,
+      "completions/min_terminated_length": 1560.625,
+      "entropy": 0.28013237891718745,
+      "epoch": 0.11009174311926606,
+      "geo": 0.8601852916181087,
+      "grad_norm": 0.322265625,
+      "learning_rate": 8.487437185929648e-06,
+      "loss": 0.0003,
+      "num_tokens": 36883412.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8154296875,
+      "reward": 0.8752201721072197,
+      "step": 312
+    },
+    {
+      "3d_bbox": 0.7732156217098236,
+      "adv": 0.00015621728016412817,
+      "bbox": 0.9900884553790092,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.25,
+      "completions/max_terminated_length": 1595.25,
+      "completions/mean_length": 1574.06640625,
+      "completions/mean_terminated_length": 1574.06640625,
+      "completions/min_length": 1556.5,
+      "completions/min_terminated_length": 1556.5,
+      "entropy": 0.2842056662775576,
+      "epoch": 0.11291460832745237,
+      "geo": 0.8583031129091978,
+      "grad_norm": 0.359375,
+      "learning_rate": 8.447236180904524e-06,
+      "loss": -0.0003,
+      "num_tokens": 37827701.0,
+      "point_2d": 0.8583984375,
+      "point_3d": 0.8681640625,
+      "reward": 0.8724666386842728,
+      "step": 320
+    },
+    {
+      "3d_bbox": 0.7427574843168259,
+      "adv": -0.00022968700068304315,
+      "bbox": 0.9458357989788055,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.25,
+      "completions/max_terminated_length": 1599.25,
+      "completions/mean_length": 1576.78125,
+      "completions/mean_terminated_length": 1576.78125,
+      "completions/min_length": 1555.625,
+      "completions/min_terminated_length": 1555.625,
+      "entropy": 0.28632913110777736,
+      "epoch": 0.11573747353563868,
+      "geo": 0.9527824521064758,
+      "grad_norm": 0.2578125,
+      "learning_rate": 8.407035175879398e-06,
+      "loss": -0.0,
+      "num_tokens": 38772853.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8623046875,
+      "reward": 0.8877245113253593,
+      "step": 328
+    },
+    {
+      "3d_bbox": 0.7519282810389996,
+      "adv": 0.0003111484527380526,
+      "bbox": 0.9336327686905861,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.0,
+      "completions/max_terminated_length": 1598.0,
+      "completions/mean_length": 1578.54296875,
+      "completions/mean_terminated_length": 1578.54296875,
+      "completions/min_length": 1560.0,
+      "completions/min_terminated_length": 1560.0,
+      "entropy": 0.2819040766917169,
+      "epoch": 0.11856033874382499,
+      "geo": 0.9259950742125511,
+      "grad_norm": 0.306640625,
+      "learning_rate": 8.366834170854273e-06,
+      "loss": 0.0202,
+      "num_tokens": 39718400.0,
+      "point_2d": 0.7451171875,
+      "point_3d": 0.748046875,
+      "reward": 0.7946812808513641,
+      "step": 336
+    },
+    {
+      "3d_bbox": 0.6988782063126564,
+      "adv": 8.153711894226534e-05,
+      "bbox": 0.9918540865182877,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1602.0,
+      "completions/max_terminated_length": 1602.0,
+      "completions/mean_length": 1578.3984375,
+      "completions/mean_terminated_length": 1578.3984375,
+      "completions/min_length": 1557.0,
+      "completions/min_terminated_length": 1557.0,
+      "entropy": 0.2852953290566802,
+      "epoch": 0.1213832039520113,
+      "geo": 0.824953593313694,
+      "grad_norm": 0.33984375,
+      "learning_rate": 8.326633165829147e-06,
+      "loss": 0.0003,
+      "num_tokens": 40663774.0,
+      "point_2d": 0.875,
+      "point_3d": 0.771484375,
+      "reward": 0.8343041799962521,
+      "step": 344
+    },
+    {
+      "3d_bbox": 0.7085578441619873,
+      "adv": 0.000511349891894497,
+      "bbox": 0.9954689294099808,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.625,
+      "completions/max_terminated_length": 1601.625,
+      "completions/mean_length": 1578.21484375,
+      "completions/mean_terminated_length": 1578.21484375,
+      "completions/min_length": 1557.5,
+      "completions/min_terminated_length": 1557.5,
+      "entropy": 0.2835490182042122,
+      "epoch": 0.12420606916019761,
+      "geo": 0.8287867829203606,
+      "grad_norm": 0.31640625,
+      "learning_rate": 8.286432160804021e-06,
+      "loss": 0.0,
+      "num_tokens": 41609261.0,
+      "point_2d": 0.875,
+      "point_3d": 0.8564453125,
+      "reward": 0.8588679879903793,
+      "step": 352
+    },
+    {
+      "3d_bbox": 0.7324622049927711,
+      "adv": -0.06052318853471661,
+      "bbox": 0.9785870239138603,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.75,
+      "completions/max_terminated_length": 1599.75,
+      "completions/mean_length": 1580.8671875,
+      "completions/mean_terminated_length": 1580.8671875,
+      "completions/min_length": 1562.875,
+      "completions/min_terminated_length": 1562.875,
+      "entropy": 0.2788377907127142,
+      "epoch": 0.12702893436838392,
+      "geo": 0.8886202499270439,
+      "grad_norm": 0.765625,
+      "learning_rate": 8.246231155778895e-06,
+      "loss": 0.0295,
+      "num_tokens": 42555355.0,
+      "point_2d": 0.984375,
+      "point_3d": 0.80078125,
+      "reward": 0.8740513622760773,
+      "step": 360
+    },
+    {
+      "3d_bbox": 0.6365351751446724,
+      "adv": 0.0003275554918218404,
+      "bbox": 0.8720735665410757,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.875,
+      "completions/max_terminated_length": 1601.875,
+      "completions/mean_length": 1577.73046875,
+      "completions/mean_terminated_length": 1577.73046875,
+      "completions/min_length": 1555.75,
+      "completions/min_terminated_length": 1555.75,
+      "entropy": 0.28446380188688636,
+      "epoch": 0.12985179957657023,
+      "geo": 0.8144352469826117,
+      "grad_norm": 0.34375,
+      "learning_rate": 8.20603015075377e-06,
+      "loss": -0.0001,
+      "num_tokens": 43500534.0,
+      "point_2d": 0.875,
+      "point_3d": 0.7666015625,
+      "reward": 0.787552572786808,
+      "step": 368
+    },
+    {
+      "3d_bbox": 0.7198036424815655,
+      "adv": -9.503715409664437e-05,
+      "bbox": 0.9939679279923439,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.75,
+      "completions/max_terminated_length": 1592.75,
+      "completions/mean_length": 1573.19140625,
+      "completions/mean_terminated_length": 1573.19140625,
+      "completions/min_length": 1557.0,
+      "completions/min_terminated_length": 1557.0,
+      "entropy": 0.28550330083817244,
+      "epoch": 0.13267466478475654,
+      "geo": 0.9650512784719467,
+      "grad_norm": 0.359375,
+      "learning_rate": 8.165829145728644e-06,
+      "loss": 0.0004,
+      "num_tokens": 44444583.0,
+      "point_2d": 0.9892578125,
+      "point_3d": 0.83984375,
+      "reward": 0.8857182934880257,
+      "step": 376
+    },
+    {
+      "3d_bbox": 0.7569727748632431,
+      "adv": 0.0004161505894444417,
+      "bbox": 0.9042535312473774,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.5,
+      "completions/max_terminated_length": 1599.5,
+      "completions/mean_length": 1577.55859375,
+      "completions/mean_terminated_length": 1577.55859375,
+      "completions/min_length": 1561.375,
+      "completions/min_terminated_length": 1561.375,
+      "entropy": 0.2907078079879284,
+      "epoch": 0.13549752999294284,
+      "geo": 0.6764855124056339,
+      "grad_norm": 0.345703125,
+      "learning_rate": 8.125628140703518e-06,
+      "loss": 0.0001,
+      "num_tokens": 45389846.0,
+      "point_2d": 0.69140625,
+      "point_3d": 0.8193359375,
+      "reward": 0.7929921373724937,
+      "step": 384
+    },
+    {
+      "3d_bbox": 0.6764840334653854,
+      "adv": -1.4531655324390158e-05,
+      "bbox": 0.8646975085139275,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1600.5,
+      "completions/max_terminated_length": 1600.5,
+      "completions/mean_length": 1579.08984375,
+      "completions/mean_terminated_length": 1579.08984375,
+      "completions/min_length": 1558.375,
+      "completions/min_terminated_length": 1558.375,
+      "entropy": 0.2863780874758959,
+      "epoch": 0.13832039520112915,
+      "geo": 0.5404472178779542,
+      "grad_norm": 0.3125,
+      "learning_rate": 8.085427135678392e-06,
+      "loss": 0.0,
+      "num_tokens": 46335629.0,
+      "point_2d": 0.5048828125,
+      "point_3d": 0.806640625,
+      "reward": 0.713176254183054,
+      "step": 392
+    },
+    {
+      "3d_bbox": 0.7255214750766754,
+      "adv": -0.0006330321593850385,
+      "bbox": 0.891293590888381,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1602.375,
+      "completions/max_terminated_length": 1602.375,
+      "completions/mean_length": 1576.9921875,
+      "completions/mean_terminated_length": 1576.9921875,
+      "completions/min_length": 1558.125,
+      "completions/min_terminated_length": 1558.125,
+      "entropy": 0.28716066014021635,
+      "epoch": 0.14114326040931546,
+      "geo": 0.9741705134510994,
+      "grad_norm": 0.314453125,
+      "learning_rate": 8.045226130653267e-06,
+      "loss": 0.0004,
+      "num_tokens": 47280507.0,
+      "point_2d": 0.8916015625,
+      "point_3d": 0.859375,
+      "reward": 0.8419479243457317,
+      "step": 400
+    },
+    {
+      "3d_bbox": 0.7029357459396124,
+      "adv": 0.0008838025460136123,
+      "bbox": 0.9950641989707947,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.125,
+      "completions/max_terminated_length": 1598.125,
+      "completions/mean_length": 1575.92578125,
+      "completions/mean_terminated_length": 1575.92578125,
+      "completions/min_length": 1557.75,
+      "completions/min_terminated_length": 1557.75,
+      "entropy": 0.2837083409540355,
+      "epoch": 0.14396612561750177,
+      "geo": 0.864502564072609,
+      "grad_norm": 0.345703125,
+      "learning_rate": 8.005025125628141e-06,
+      "loss": -0.0003,
+      "num_tokens": 48225240.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8310546875,
+      "reward": 0.8822636753320694,
+      "step": 408
+    },
+    {
+      "3d_bbox": 0.6913789249956608,
+      "adv": -0.0002196824352722615,
+      "bbox": 0.9765095412731171,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.125,
+      "completions/max_terminated_length": 1598.125,
+      "completions/mean_length": 1579.28515625,
+      "completions/mean_terminated_length": 1579.28515625,
+      "completions/min_length": 1559.25,
+      "completions/min_terminated_length": 1559.25,
+      "entropy": 0.28565866220742464,
+      "epoch": 0.14678899082568808,
+      "geo": 0.8599895760416985,
+      "grad_norm": 0.30859375,
+      "learning_rate": 7.964824120603015e-06,
+      "loss": 0.0001,
+      "num_tokens": 49171057.0,
+      "point_2d": 1.0,
+      "point_3d": 0.904296875,
+      "reward": 0.8930463194847107,
+      "step": 416
+    },
+    {
+      "3d_bbox": 0.6788053698837757,
+      "adv": -0.00020617676273104735,
+      "bbox": 0.9748031497001648,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.25,
+      "completions/max_terminated_length": 1595.25,
+      "completions/mean_length": 1577.51953125,
+      "completions/mean_terminated_length": 1577.51953125,
+      "completions/min_length": 1559.625,
+      "completions/min_terminated_length": 1559.625,
+      "entropy": 0.28440852370113134,
+      "epoch": 0.1496118560338744,
+      "geo": 0.822394497692585,
+      "grad_norm": 0.40234375,
+      "learning_rate": 7.92462311557789e-06,
+      "loss": 0.0006,
+      "num_tokens": 50116302.0,
+      "point_2d": 0.875,
+      "point_3d": 0.8310546875,
+      "reward": 0.8399158045649529,
+      "step": 424
+    },
+    {
+      "3d_bbox": 0.6794650703668594,
+      "adv": 0.0011559298304746335,
+      "bbox": 0.9097459949553013,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.625,
+      "completions/max_terminated_length": 1596.625,
+      "completions/mean_length": 1573.80859375,
+      "completions/mean_terminated_length": 1573.80859375,
+      "completions/min_length": 1552.625,
+      "completions/min_terminated_length": 1552.625,
+      "entropy": 0.28376144589856267,
+      "epoch": 0.1524347212420607,
+      "geo": 0.7893084269016981,
+      "grad_norm": 0.3203125,
+      "learning_rate": 7.884422110552764e-06,
+      "loss": 0.0,
+      "num_tokens": 51060365.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8681640625,
+      "reward": 0.8643437847495079,
+      "step": 432
+    },
+    {
+      "3d_bbox": 0.7942164540290833,
+      "adv": -3.183347871527076e-05,
+      "bbox": 0.9926317036151886,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.75,
+      "completions/max_terminated_length": 1598.75,
+      "completions/mean_length": 1577.796875,
+      "completions/mean_terminated_length": 1577.796875,
+      "completions/min_length": 1553.125,
+      "completions/min_terminated_length": 1553.125,
+      "entropy": 0.2821194725111127,
+      "epoch": 0.155257586450247,
+      "geo": 0.9503362327814102,
+      "grad_norm": 0.3984375,
+      "learning_rate": 7.844221105527638e-06,
+      "loss": 0.0002,
+      "num_tokens": 52005633.0,
+      "point_2d": 1.0,
+      "point_3d": 0.869140625,
+      "reward": 0.9139972031116486,
+      "step": 440
+    },
+    {
+      "3d_bbox": 0.6752692721784115,
+      "adv": -9.695853077573702e-05,
+      "bbox": 0.982423759996891,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.25,
+      "completions/max_terminated_length": 1596.25,
+      "completions/mean_length": 1574.5546875,
+      "completions/mean_terminated_length": 1574.5546875,
+      "completions/min_length": 1555.125,
+      "completions/min_terminated_length": 1555.125,
+      "entropy": 0.2852726257406175,
+      "epoch": 0.15808045165843332,
+      "geo": 0.5512092066928744,
+      "grad_norm": 0.322265625,
+      "learning_rate": 7.804020100502512e-06,
+      "loss": -0.0004,
+      "num_tokens": 52949951.0,
+      "point_2d": 0.75,
+      "point_3d": 0.7998046875,
+      "reward": 0.801874428987503,
+      "step": 448
+    },
+    {
+      "3d_bbox": 0.7081510461866856,
+      "adv": 0.0003901398231391795,
+      "bbox": 0.9892760217189789,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.875,
+      "completions/max_terminated_length": 1596.875,
+      "completions/mean_length": 1573.0390625,
+      "completions/mean_terminated_length": 1573.0390625,
+      "completions/min_length": 1551.875,
+      "completions/min_terminated_length": 1551.875,
+      "entropy": 0.2837893716059625,
+      "epoch": 0.16090331686661963,
+      "geo": 0.9603376165032387,
+      "grad_norm": 0.3125,
+      "learning_rate": 7.763819095477387e-06,
+      "loss": -0.0007,
+      "num_tokens": 53894057.0,
+      "point_2d": 1.0,
+      "point_3d": 0.794921875,
+      "reward": 0.8730872422456741,
+      "step": 456
+    },
+    {
+      "3d_bbox": 0.7504811957478523,
+      "adv": 0.0007141208916436881,
+      "bbox": 0.9912561923265457,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.75,
+      "completions/max_terminated_length": 1594.75,
+      "completions/mean_length": 1576.1640625,
+      "completions/mean_terminated_length": 1576.1640625,
+      "completions/min_length": 1555.75,
+      "completions/min_terminated_length": 1555.75,
+      "entropy": 0.28930201614275575,
+      "epoch": 0.16372618207480594,
+      "geo": 0.8148366361856461,
+      "grad_norm": 0.275390625,
+      "learning_rate": 7.723618090452262e-06,
+      "loss": 0.0003,
+      "num_tokens": 54839099.0,
+      "point_2d": 0.875,
+      "point_3d": 0.8896484375,
+      "reward": 0.8765964508056641,
+      "step": 464
+    },
+    {
+      "3d_bbox": 0.819125022739172,
+      "adv": 0.00010633431156747974,
+      "bbox": 0.9906315729022026,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1603.0,
+      "completions/max_terminated_length": 1603.0,
+      "completions/mean_length": 1578.3203125,
+      "completions/mean_terminated_length": 1578.3203125,
+      "completions/min_length": 1556.375,
+      "completions/min_terminated_length": 1556.375,
+      "entropy": 0.28396829310804605,
+      "epoch": 0.16654904728299225,
+      "geo": 0.8597357729449868,
+      "grad_norm": 0.30078125,
+      "learning_rate": 7.683417085427137e-06,
+      "loss": 0.0004,
+      "num_tokens": 55784557.0,
+      "point_2d": 0.8916015625,
+      "point_3d": 0.9345703125,
+      "reward": 0.9089821055531502,
+      "step": 472
+    },
+    {
+      "3d_bbox": 0.7331585213541985,
+      "adv": -0.00033648866519797593,
+      "bbox": 0.8809062782675028,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.625,
+      "completions/max_terminated_length": 1594.625,
+      "completions/mean_length": 1576.11328125,
+      "completions/mean_terminated_length": 1576.11328125,
+      "completions/min_length": 1559.25,
+      "completions/min_terminated_length": 1559.25,
+      "entropy": 0.28669640561565757,
+      "epoch": 0.16937191249117856,
+      "geo": 0.9506581574678421,
+      "grad_norm": 0.2890625,
+      "learning_rate": 7.643216080402011e-06,
+      "loss": 0.0132,
+      "num_tokens": 56729514.0,
+      "point_2d": 0.8759765625,
+      "point_3d": 0.880859375,
+      "reward": 0.8427251875400543,
+      "step": 480
+    },
+    {
+      "3d_bbox": 0.7568283602595329,
+      "adv": -0.0008519564944435842,
+      "bbox": 0.9879242181777954,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1600.875,
+      "completions/max_terminated_length": 1600.875,
+      "completions/mean_length": 1574.87890625,
+      "completions/mean_terminated_length": 1574.87890625,
+      "completions/min_length": 1554.25,
+      "completions/min_terminated_length": 1554.25,
+      "entropy": 0.28737954096868634,
+      "epoch": 0.17219477769936486,
+      "geo": 0.7571075335144997,
+      "grad_norm": 0.400390625,
+      "learning_rate": 7.603015075376885e-06,
+      "loss": 0.0105,
+      "num_tokens": 57674099.0,
+      "point_2d": 0.7998046875,
+      "point_3d": 0.9228515625,
+      "reward": 0.8668522015213966,
+      "step": 488
+    },
+    {
+      "3d_bbox": 0.7658256888389587,
+      "adv": 0.0011402575073589105,
+      "bbox": 0.9932135492563248,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.25,
+      "completions/max_terminated_length": 1596.25,
+      "completions/mean_length": 1574.265625,
+      "completions/mean_terminated_length": 1574.265625,
+      "completions/min_length": 1554.375,
+      "completions/min_terminated_length": 1554.375,
+      "entropy": 0.2865150640718639,
+      "epoch": 0.17501764290755117,
+      "geo": 0.8641151636838913,
+      "grad_norm": 0.37890625,
+      "learning_rate": 7.5628140703517595e-06,
+      "loss": -0.0004,
+      "num_tokens": 58618575.0,
+      "point_2d": 0.91015625,
+      "point_3d": 0.904296875,
+      "reward": 0.8933730870485306,
+      "step": 496
+    },
+    {
+      "3d_bbox": 0.7145381905138493,
+      "adv": 0.0003104716488451231,
+      "bbox": 0.9926531538367271,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.75,
+      "completions/max_terminated_length": 1594.75,
+      "completions/mean_length": 1573.12890625,
+      "completions/mean_terminated_length": 1573.12890625,
+      "completions/min_length": 1553.375,
+      "completions/min_terminated_length": 1553.375,
+      "entropy": 0.28603550279513,
+      "epoch": 0.17784050811573748,
+      "geo": 0.7219396531581879,
+      "grad_norm": 0.310546875,
+      "learning_rate": 7.522613065326634e-06,
+      "loss": -0.0,
+      "num_tokens": 59562552.0,
+      "point_2d": 0.7509765625,
+      "point_3d": 0.9365234375,
+      "reward": 0.8486728444695473,
+      "step": 504
+    },
+    {
+      "3d_bbox": 0.7242622300982475,
+      "adv": -0.0005710294717573561,
+      "bbox": 0.9246156923472881,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1591.75,
+      "completions/max_terminated_length": 1591.75,
+      "completions/mean_length": 1571.015625,
+      "completions/mean_terminated_length": 1571.015625,
+      "completions/min_length": 1551.625,
+      "completions/min_terminated_length": 1551.625,
+      "entropy": 0.2915982282720506,
+      "epoch": 0.1806633733239238,
+      "geo": 0.7993985265493393,
+      "grad_norm": 0.408203125,
+      "learning_rate": 7.482412060301508e-06,
+      "loss": 0.0001,
+      "num_tokens": 60505948.0,
+      "point_2d": 0.7919921875,
+      "point_3d": 0.8671875,
+      "reward": 0.8270144015550613,
+      "step": 512
+    },
+    {
+      "3d_bbox": 0.7831042557954788,
+      "adv": -0.000159245344093506,
+      "bbox": 0.9960433393716812,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.375,
+      "completions/max_terminated_length": 1592.375,
+      "completions/mean_length": 1574.84375,
+      "completions/mean_terminated_length": 1574.84375,
+      "completions/min_length": 1559.125,
+      "completions/min_terminated_length": 1559.125,
+      "entropy": 0.2845937805250287,
+      "epoch": 0.1834862385321101,
+      "geo": 0.9170719608664513,
+      "grad_norm": 0.345703125,
+      "learning_rate": 7.442211055276382e-06,
+      "loss": -0.0,
+      "num_tokens": 61450564.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8408203125,
+      "reward": 0.90499197691679,
+      "step": 520
+    },
+    {
+      "3d_bbox": 0.7821980714797974,
+      "adv": -0.000531579293010509,
+      "bbox": 0.9920122027397156,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.375,
+      "completions/max_terminated_length": 1595.375,
+      "completions/mean_length": 1571.765625,
+      "completions/mean_terminated_length": 1571.765625,
+      "completions/min_length": 1551.5,
+      "completions/min_terminated_length": 1551.5,
+      "entropy": 0.2882155287079513,
+      "epoch": 0.1863091037402964,
+      "geo": 0.901853509247303,
+      "grad_norm": 0.365234375,
+      "learning_rate": 7.402010050251257e-06,
+      "loss": 0.0002,
+      "num_tokens": 62394336.0,
+      "point_2d": 0.888671875,
+      "point_3d": 0.9052734375,
+      "reward": 0.8920388966798782,
+      "step": 528
+    },
+    {
+      "3d_bbox": 0.7537402808666229,
+      "adv": -0.0004438982141437009,
+      "bbox": 0.9946581870317459,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.375,
+      "completions/max_terminated_length": 1594.375,
+      "completions/mean_length": 1574.3203125,
+      "completions/mean_terminated_length": 1574.3203125,
+      "completions/min_length": 1554.75,
+      "completions/min_terminated_length": 1554.75,
+      "entropy": 0.2886669235303998,
+      "epoch": 0.18913196894848272,
+      "geo": 0.8310712724924088,
+      "grad_norm": 0.36328125,
+      "learning_rate": 7.361809045226132e-06,
+      "loss": 0.0002,
+      "num_tokens": 63338834.0,
+      "point_2d": 0.873046875,
+      "point_3d": 0.904296875,
+      "reward": 0.8814355507493019,
+      "step": 536
+    },
+    {
+      "3d_bbox": 0.6960996612906456,
+      "adv": 0.0006180387658787367,
+      "bbox": 0.869446316966787,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1597.25,
+      "completions/max_terminated_length": 1597.25,
+      "completions/mean_length": 1571.88671875,
+      "completions/mean_terminated_length": 1571.88671875,
+      "completions/min_length": 1552.25,
+      "completions/min_terminated_length": 1552.25,
+      "entropy": 0.2887332271784544,
+      "epoch": 0.19195483415666903,
+      "geo": 0.826970137655735,
+      "grad_norm": 0.267578125,
+      "learning_rate": 7.321608040201006e-06,
+      "loss": -0.0,
+      "num_tokens": 64282685.0,
+      "point_2d": 0.875,
+      "point_3d": 0.8662109375,
+      "reward": 0.8266892246901989,
+      "step": 544
+    },
+    {
+      "3d_bbox": 0.8654715269804001,
+      "adv": -0.0008252508814621251,
+      "bbox": 0.9932455867528915,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.125,
+      "completions/max_terminated_length": 1596.125,
+      "completions/mean_length": 1572.1953125,
+      "completions/mean_terminated_length": 1572.1953125,
+      "completions/min_length": 1553.625,
+      "completions/min_terminated_length": 1553.625,
+      "entropy": 0.2873402568511665,
+      "epoch": 0.19477769936485534,
+      "geo": 0.872327484190464,
+      "grad_norm": 0.2890625,
+      "learning_rate": 7.28140703517588e-06,
+      "loss": 0.0004,
+      "num_tokens": 65226639.0,
+      "point_2d": 0.8935546875,
+      "point_3d": 0.9482421875,
+      "reward": 0.9251285046339035,
+      "step": 552
+    },
+    {
+      "3d_bbox": 0.7851101011037827,
+      "adv": -0.0011301148442726117,
+      "bbox": 0.9379025921225548,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1588.5,
+      "completions/max_terminated_length": 1588.5,
+      "completions/mean_length": 1571.4765625,
+      "completions/mean_terminated_length": 1571.4765625,
+      "completions/min_length": 1555.0,
+      "completions/min_terminated_length": 1555.0,
+      "entropy": 0.2852159277535975,
+      "epoch": 0.19760056457304165,
+      "geo": 0.690475556999445,
+      "grad_norm": 0.3125,
+      "learning_rate": 7.241206030150754e-06,
+      "loss": 0.0006,
+      "num_tokens": 66170305.0,
+      "point_2d": 0.625,
+      "point_3d": 0.837890625,
+      "reward": 0.7964758351445198,
+      "step": 560
+    },
+    {
+      "3d_bbox": 0.7313521131873131,
+      "adv": -0.000422738248744281,
+      "bbox": 0.8720479012117721,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.875,
+      "completions/max_terminated_length": 1601.875,
+      "completions/mean_length": 1574.91015625,
+      "completions/mean_terminated_length": 1574.91015625,
+      "completions/min_length": 1551.25,
+      "completions/min_terminated_length": 1551.25,
+      "entropy": 0.2880417066626251,
+      "epoch": 0.20042342978122796,
+      "geo": 0.8383861493784934,
+      "grad_norm": 0.412109375,
+      "learning_rate": 7.2010050251256295e-06,
+      "loss": 0.0004,
+      "num_tokens": 67114842.0,
+      "point_2d": 0.8740234375,
+      "point_3d": 0.84765625,
+      "reward": 0.8312699124217033,
+      "step": 568
+    },
+    {
+      "3d_bbox": 0.6841592267155647,
+      "adv": -0.0008141829575833981,
+      "bbox": 0.9068506434559822,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1597.5,
+      "completions/max_terminated_length": 1597.5,
+      "completions/mean_length": 1571.79296875,
+      "completions/mean_terminated_length": 1571.79296875,
+      "completions/min_length": 1546.75,
+      "completions/min_terminated_length": 1546.75,
+      "entropy": 0.2889366364106536,
+      "epoch": 0.20324629498941427,
+      "geo": 0.7308010272681713,
+      "grad_norm": 0.2578125,
+      "learning_rate": 7.160804020100504e-06,
+      "loss": 0.0002,
+      "num_tokens": 68058501.0,
+      "point_2d": 0.8525390625,
+      "point_3d": 0.794921875,
+      "reward": 0.8096176832914352,
+      "step": 576
+    },
+    {
+      "3d_bbox": 0.5905975513160229,
+      "adv": -0.00022596626513404772,
+      "bbox": 0.9184868931770325,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1597.5,
+      "completions/max_terminated_length": 1597.5,
+      "completions/mean_length": 1572.45703125,
+      "completions/mean_terminated_length": 1572.45703125,
+      "completions/min_length": 1551.125,
+      "completions/min_terminated_length": 1551.125,
+      "entropy": 0.2837987565435469,
+      "epoch": 0.20606916019760058,
+      "geo": 0.8267702832818031,
+      "grad_norm": 0.314453125,
+      "learning_rate": 7.120603015075378e-06,
+      "loss": -0.0003,
+      "num_tokens": 69002250.0,
+      "point_2d": 0.8681640625,
+      "point_3d": 0.7445197626948357,
+      "reward": 0.7804420664906502,
+      "step": 584
+    },
+    {
+      "3d_bbox": 0.6897582598030567,
+      "adv": -0.0004831653022847604,
+      "bbox": 0.9742325842380524,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.25,
+      "completions/max_terminated_length": 1595.25,
+      "completions/mean_length": 1572.9296875,
+      "completions/mean_terminated_length": 1572.9296875,
+      "completions/min_length": 1553.5,
+      "completions/min_terminated_length": 1553.5,
+      "entropy": 0.2835669592022896,
+      "epoch": 0.20889202540578689,
+      "geo": 0.8361655548214912,
+      "grad_norm": 0.330078125,
+      "learning_rate": 7.080402010050251e-06,
+      "loss": 0.0004,
+      "num_tokens": 69946400.0,
+      "point_2d": 0.875,
+      "point_3d": 0.8948184698820114,
+      "reward": 0.8584523350000381,
+      "step": 592
+    },
+    {
+      "3d_bbox": 0.785271767526865,
+      "adv": -0.00028875649149995297,
+      "bbox": 0.9752914533019066,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.0,
+      "completions/max_terminated_length": 1599.0,
+      "completions/mean_length": 1572.3671875,
+      "completions/mean_terminated_length": 1572.3671875,
+      "completions/min_length": 1549.875,
+      "completions/min_terminated_length": 1549.875,
+      "entropy": 0.29118923749774694,
+      "epoch": 0.2117148906139732,
+      "geo": 0.9338327869772911,
+      "grad_norm": 0.3671875,
+      "learning_rate": 7.040201005025126e-06,
+      "loss": 0.0007,
+      "num_tokens": 70890342.0,
+      "point_2d": 0.998046875,
+      "point_3d": 0.94140625,
+      "reward": 0.9250040724873543,
+      "step": 600
+    },
+    {
+      "3d_bbox": 0.8766504302620888,
+      "adv": -0.002171848598663928,
+      "bbox": 0.9872715324163437,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.125,
+      "completions/max_terminated_length": 1601.125,
+      "completions/mean_length": 1576.87890625,
+      "completions/mean_terminated_length": 1576.87890625,
+      "completions/min_length": 1556.375,
+      "completions/min_terminated_length": 1556.375,
+      "entropy": 0.28933983854949474,
+      "epoch": 0.2145377558221595,
+      "geo": 0.845995269715786,
+      "grad_norm": 0.33203125,
+      "learning_rate": 7e-06,
+      "loss": 0.0003,
+      "num_tokens": 71835359.0,
+      "point_2d": 0.875,
+      "point_3d": 0.9454848319292068,
+      "reward": 0.9211016893386841,
+      "step": 608
+    },
+    {
+      "3d_bbox": 0.8059130385518074,
+      "adv": 0.0005476666337926872,
+      "bbox": 0.9907352328300476,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1593.625,
+      "completions/max_terminated_length": 1593.625,
+      "completions/mean_length": 1570.9921875,
+      "completions/mean_terminated_length": 1570.9921875,
+      "completions/min_length": 1554.125,
+      "completions/min_terminated_length": 1554.125,
+      "entropy": 0.28401868836954236,
+      "epoch": 0.2173606210303458,
+      "geo": 0.8720198646187782,
+      "grad_norm": 0.384765625,
+      "learning_rate": 6.959798994974874e-06,
+      "loss": -0.0004,
+      "num_tokens": 72778933.0,
+      "point_2d": 0.9248046875,
+      "point_3d": 0.9135454967617989,
+      "reward": 0.9087496176362038,
+      "step": 616
+    },
+    {
+      "3d_bbox": 0.8549554497003555,
+      "adv": -0.0005420902471087174,
+      "bbox": 0.9758080616593361,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1593.0,
+      "completions/max_terminated_length": 1593.0,
+      "completions/mean_length": 1570.1640625,
+      "completions/mean_terminated_length": 1570.1640625,
+      "completions/min_length": 1549.0,
+      "completions/min_terminated_length": 1549.0,
+      "entropy": 0.29045473085716367,
+      "epoch": 0.22018348623853212,
+      "geo": 0.8916067183017731,
+      "grad_norm": 0.3828125,
+      "learning_rate": 6.919597989949749e-06,
+      "loss": 0.0156,
+      "num_tokens": 73722191.0,
+      "point_2d": 1.0,
+      "point_3d": 0.9639246314764023,
+      "reward": 0.9486720338463783,
+      "step": 624
+    },
+    {
+      "3d_bbox": 0.7975610345602036,
+      "adv": 0.0014872762099003012,
+      "bbox": 0.9193930514156818,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1600.625,
+      "completions/max_terminated_length": 1600.625,
+      "completions/mean_length": 1573.29296875,
+      "completions/mean_terminated_length": 1573.29296875,
+      "completions/min_length": 1553.0,
+      "completions/min_terminated_length": 1553.0,
+      "entropy": 0.28542913123965263,
+      "epoch": 0.22300635144671843,
+      "geo": 0.936805784702301,
+      "grad_norm": 0.2734375,
+      "learning_rate": 6.8793969849246235e-06,
+      "loss": 0.0152,
+      "num_tokens": 74666354.0,
+      "point_2d": 0.9990234375,
+      "point_3d": 0.93359375,
+      "reward": 0.9123928248882294,
+      "step": 632
+    },
+    {
+      "3d_bbox": 0.6216618679463863,
+      "adv": -0.0013801682944176719,
+      "bbox": 0.9884917959570885,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.375,
+      "completions/max_terminated_length": 1601.375,
+      "completions/mean_length": 1577.00390625,
+      "completions/mean_terminated_length": 1577.00390625,
+      "completions/min_length": 1554.0,
+      "completions/min_terminated_length": 1554.0,
+      "entropy": 0.28678226098418236,
+      "epoch": 0.22582921665490474,
+      "geo": 0.9096385985612869,
+      "grad_norm": 0.42578125,
+      "learning_rate": 6.839195979899498e-06,
+      "loss": 0.0005,
+      "num_tokens": 75611339.0,
+      "point_2d": 0.8291015625,
+      "point_3d": 0.7913028448820114,
+      "reward": 0.8076395243406296,
+      "step": 640
+    },
+    {
+      "3d_bbox": 0.7582510337233543,
+      "adv": -0.00019565537664334443,
+      "bbox": 0.9911796301603317,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.625,
+      "completions/max_terminated_length": 1596.625,
+      "completions/mean_length": 1575.71484375,
+      "completions/mean_terminated_length": 1575.71484375,
+      "completions/min_length": 1556.25,
+      "completions/min_terminated_length": 1556.25,
+      "entropy": 0.2863654876127839,
+      "epoch": 0.22865208186309105,
+      "geo": 0.9025658443570137,
+      "grad_norm": 0.451171875,
+      "learning_rate": 6.798994974874372e-06,
+      "loss": -0.0001,
+      "num_tokens": 76556042.0,
+      "point_2d": 0.95703125,
+      "point_3d": 0.8994140625,
+      "reward": 0.9014689922332764,
+      "step": 648
+    },
+    {
+      "3d_bbox": 0.753610834479332,
+      "adv": 0.00039293405916396296,
+      "bbox": 0.9021090529859066,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.625,
+      "completions/max_terminated_length": 1592.625,
+      "completions/mean_length": 1572.203125,
+      "completions/mean_terminated_length": 1572.203125,
+      "completions/min_length": 1552.25,
+      "completions/min_terminated_length": 1552.25,
+      "entropy": 0.286099414806813,
+      "epoch": 0.23147494707127736,
+      "geo": 0.9557241797447205,
+      "grad_norm": 0.455078125,
+      "learning_rate": 6.758793969849246e-06,
+      "loss": 0.0002,
+      "num_tokens": 77499886.0,
+      "point_2d": 1.0,
+      "point_3d": 0.91015625,
+      "reward": 0.8914690390229225,
+      "step": 656
+    },
+    {
+      "3d_bbox": 0.831758588552475,
+      "adv": -7.158980952226557e-05,
+      "bbox": 0.9794979766011238,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.5,
+      "completions/max_terminated_length": 1595.5,
+      "completions/mean_length": 1574.5,
+      "completions/mean_terminated_length": 1574.5,
+      "completions/min_length": 1558.75,
+      "completions/min_terminated_length": 1558.75,
+      "entropy": 0.2819104827940464,
+      "epoch": 0.23429781227946367,
+      "geo": 0.8940749987959862,
+      "grad_norm": 0.294921875,
+      "learning_rate": 6.718592964824121e-06,
+      "loss": 0.0001,
+      "num_tokens": 78444526.0,
+      "point_2d": 1.0,
+      "point_3d": 0.939453125,
+      "reward": 0.9376774355769157,
+      "step": 664
+    },
+    {
+      "3d_bbox": 0.8210256397724152,
+      "adv": 0.000540634344361024,
+      "bbox": 0.9576755836606026,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.125,
+      "completions/max_terminated_length": 1594.125,
+      "completions/mean_length": 1573.80859375,
+      "completions/mean_terminated_length": 1573.80859375,
+      "completions/min_length": 1553.25,
+      "completions/min_terminated_length": 1553.25,
+      "entropy": 0.2863646815530956,
+      "epoch": 0.23712067748764998,
+      "geo": 0.9599328562617302,
+      "grad_norm": 0.28125,
+      "learning_rate": 6.678391959798996e-06,
+      "loss": -0.0001,
+      "num_tokens": 79388949.0,
+      "point_2d": 1.0,
+      "point_3d": 0.92578125,
+      "reward": 0.9261206015944481,
+      "step": 672
+    },
+    {
+      "3d_bbox": 0.8162731677293777,
+      "adv": -0.0006006620242260396,
+      "bbox": 0.9639560282230377,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1593.75,
+      "completions/max_terminated_length": 1593.75,
+      "completions/mean_length": 1573.62890625,
+      "completions/mean_terminated_length": 1573.62890625,
+      "completions/min_length": 1557.125,
+      "completions/min_terminated_length": 1557.125,
+      "entropy": 0.2868205299600959,
+      "epoch": 0.2399435426958363,
+      "geo": 0.9127583503723145,
+      "grad_norm": 0.44921875,
+      "learning_rate": 6.63819095477387e-06,
+      "loss": -0.0,
+      "num_tokens": 80333190.0,
+      "point_2d": 1.0,
+      "point_3d": 0.9267578125,
+      "reward": 0.926746778190136,
+      "step": 680
+    },
+    {
+      "3d_bbox": 0.7580128684639931,
+      "adv": 0.0009847661577850886,
+      "bbox": 0.9946200400590897,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.0,
+      "completions/max_terminated_length": 1598.0,
+      "completions/mean_length": 1572.2109375,
+      "completions/mean_terminated_length": 1572.2109375,
+      "completions/min_length": 1552.0,
+      "completions/min_terminated_length": 1552.0,
+      "entropy": 0.286160247400403,
+      "epoch": 0.2427664079040226,
+      "geo": 0.9191197454929352,
+      "grad_norm": 0.361328125,
+      "learning_rate": 6.597989949748744e-06,
+      "loss": 0.0001,
+      "num_tokens": 81277036.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8818359375,
+      "reward": 0.9086171984672546,
+      "step": 688
+    },
+    {
+      "3d_bbox": 0.729152224957943,
+      "adv": -0.0009101523137360346,
+      "bbox": 0.9927223324775696,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.875,
+      "completions/max_terminated_length": 1592.875,
+      "completions/mean_length": 1571.8515625,
+      "completions/mean_terminated_length": 1571.8515625,
+      "completions/min_length": 1553.375,
+      "completions/min_terminated_length": 1553.375,
+      "entropy": 0.28340971609577537,
+      "epoch": 0.2455892731122089,
+      "geo": 0.9836970791220665,
+      "grad_norm": 0.34765625,
+      "learning_rate": 6.557788944723618e-06,
+      "loss": 0.0005,
+      "num_tokens": 82220822.0,
+      "point_2d": 1.0,
+      "point_3d": 0.9014246314764023,
+      "reward": 0.9058247655630112,
+      "step": 696
+    },
+    {
+      "3d_bbox": 0.7259362824261189,
+      "adv": 0.0003162266129947966,
+      "bbox": 0.922937985509634,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.0,
+      "completions/max_terminated_length": 1596.0,
+      "completions/mean_length": 1575.2265625,
+      "completions/mean_terminated_length": 1575.2265625,
+      "completions/min_length": 1555.625,
+      "completions/min_terminated_length": 1555.625,
+      "entropy": 0.286905855871737,
+      "epoch": 0.24841213832039521,
+      "geo": 0.8376269713044167,
+      "grad_norm": 0.30078125,
+      "learning_rate": 6.5175879396984935e-06,
+      "loss": 0.0,
+      "num_tokens": 83165600.0,
+      "point_2d": 0.9541015625,
+      "point_3d": 0.8759765625,
+      "reward": 0.8697381019592285,
+      "step": 704
+    },
+    {
+      "3d_bbox": 0.7864346951246262,
+      "adv": -0.0007161469466154813,
+      "bbox": 0.9958559349179268,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.0,
+      "completions/max_terminated_length": 1595.0,
+      "completions/mean_length": 1572.4921875,
+      "completions/mean_terminated_length": 1572.4921875,
+      "completions/min_length": 1545.125,
+      "completions/min_terminated_length": 1545.125,
+      "entropy": 0.2886675829067826,
+      "epoch": 0.2512350035285815,
+      "geo": 0.8416303154081106,
+      "grad_norm": 0.333984375,
+      "learning_rate": 6.477386934673368e-06,
+      "loss": 0.0008,
+      "num_tokens": 84109606.0,
+      "point_2d": 0.75,
+      "point_3d": 0.947265625,
+      "reward": 0.8698890581727028,
+      "step": 712
+    },
+    {
+      "3d_bbox": 0.657585222274065,
+      "adv": -0.0004964938416378573,
+      "bbox": 0.9940767288208008,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1589.75,
+      "completions/max_terminated_length": 1589.75,
+      "completions/mean_length": 1571.515625,
+      "completions/mean_terminated_length": 1571.515625,
+      "completions/min_length": 1553.5,
+      "completions/min_terminated_length": 1553.5,
+      "entropy": 0.2854380081407726,
+      "epoch": 0.25405786873676783,
+      "geo": 0.8985359221696854,
+      "grad_norm": 0.275390625,
+      "learning_rate": 6.437185929648242e-06,
+      "loss": 0.0,
+      "num_tokens": 85053346.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8623046875,
+      "reward": 0.8784916624426842,
+      "step": 720
+    },
+    {
+      "3d_bbox": 0.8299823924899101,
+      "adv": -0.0006786447083868552,
+      "bbox": 0.9957724660634995,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.75,
+      "completions/max_terminated_length": 1599.75,
+      "completions/mean_length": 1575.05078125,
+      "completions/mean_terminated_length": 1575.05078125,
+      "completions/min_length": 1553.625,
+      "completions/min_terminated_length": 1553.625,
+      "entropy": 0.2834048862569034,
+      "epoch": 0.25688073394495414,
+      "geo": 0.7720311819575727,
+      "grad_norm": 0.353515625,
+      "learning_rate": 6.396984924623116e-06,
+      "loss": 0.0008,
+      "num_tokens": 85997743.0,
+      "point_2d": 0.875,
+      "point_3d": 0.8583984375,
+      "reward": 0.8897883221507072,
+      "step": 728
+    },
+    {
+      "3d_bbox": 0.814798042178154,
+      "adv": -0.00043147145015609567,
+      "bbox": 0.994331993162632,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1603.0,
+      "completions/max_terminated_length": 1603.0,
+      "completions/mean_length": 1578.44921875,
+      "completions/mean_terminated_length": 1578.44921875,
+      "completions/min_length": 1556.25,
+      "completions/min_terminated_length": 1556.25,
+      "entropy": 0.2801376706920564,
+      "epoch": 0.25970359915314045,
+      "geo": 0.9146922156214714,
+      "grad_norm": 0.25,
+      "learning_rate": 6.3567839195979905e-06,
+      "loss": 0.0001,
+      "num_tokens": 86943082.0,
+      "point_2d": 0.8798828125,
+      "point_3d": 0.94921875,
+      "reward": 0.909557893872261,
+      "step": 736
+    },
+    {
+      "3d_bbox": 0.7457513771951199,
+      "adv": -1.4886696476423822e-05,
+      "bbox": 0.9952592924237251,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.875,
+      "completions/max_terminated_length": 1592.875,
+      "completions/mean_length": 1572.21484375,
+      "completions/mean_terminated_length": 1572.21484375,
+      "completions/min_length": 1553.625,
+      "completions/min_terminated_length": 1553.625,
+      "entropy": 0.28585455380380154,
+      "epoch": 0.26252646436132676,
+      "geo": 0.9565036073327065,
+      "grad_norm": 0.345703125,
+      "learning_rate": 6.316582914572866e-06,
+      "loss": -0.0003,
+      "num_tokens": 87886913.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8095703125,
+      "reward": 0.8876452520489693,
+      "step": 744
+    },
+    {
+      "3d_bbox": 0.6948670484125614,
+      "adv": -0.00016846299786266172,
+      "bbox": 0.8553073853254318,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.375,
+      "completions/max_terminated_length": 1594.375,
+      "completions/mean_length": 1571.93359375,
+      "completions/mean_terminated_length": 1571.93359375,
+      "completions/min_length": 1552.0,
+      "completions/min_terminated_length": 1552.0,
+      "entropy": 0.28335432009771466,
+      "epoch": 0.26534932956951307,
+      "geo": 0.7740265503525734,
+      "grad_norm": 0.421875,
+      "learning_rate": 6.27638190954774e-06,
+      "loss": 0.0162,
+      "num_tokens": 88830776.0,
+      "point_2d": 0.8701171875,
+      "point_3d": 0.9082605689764023,
+      "reward": 0.8321380317211151,
+      "step": 752
+    },
+    {
+      "3d_bbox": 0.7499241754412651,
+      "adv": 0.0011103415663455962,
+      "bbox": 0.9951865822076797,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.75,
+      "completions/max_terminated_length": 1592.75,
+      "completions/mean_length": 1570.93359375,
+      "completions/mean_terminated_length": 1570.93359375,
+      "completions/min_length": 1552.25,
+      "completions/min_terminated_length": 1552.25,
+      "entropy": 0.28731182869523764,
+      "epoch": 0.2681721947776994,
+      "geo": 0.620879240334034,
+      "grad_norm": 0.2197265625,
+      "learning_rate": 6.236180904522614e-06,
+      "loss": 0.0001,
+      "num_tokens": 89774255.0,
+      "point_2d": 0.7119140625,
+      "point_3d": 0.8916015625,
+      "reward": 0.8371566012501717,
+      "step": 760
+    },
+    {
+      "3d_bbox": 0.7572767026722431,
+      "adv": 0.0006080263119656593,
+      "bbox": 0.9954995736479759,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1589.0,
+      "completions/max_terminated_length": 1589.0,
+      "completions/mean_length": 1569.015625,
+      "completions/mean_terminated_length": 1569.015625,
+      "completions/min_length": 1549.75,
+      "completions/min_terminated_length": 1549.75,
+      "entropy": 0.2817346737720072,
+      "epoch": 0.2709950599858857,
+      "geo": 0.8461784943938255,
+      "grad_norm": 0.41015625,
+      "learning_rate": 6.195979899497488e-06,
+      "loss": -0.0,
+      "num_tokens": 90717243.0,
+      "point_2d": 1.0,
+      "point_3d": 0.892578125,
+      "reward": 0.9113386049866676,
+      "step": 768
+    },
+    {
+      "3d_bbox": 0.8124237954616547,
+      "adv": -0.00010392009426141158,
+      "bbox": 0.9695260003209114,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1590.0,
+      "completions/max_terminated_length": 1590.0,
+      "completions/mean_length": 1571.890625,
+      "completions/mean_terminated_length": 1571.890625,
+      "completions/min_length": 1553.75,
+      "completions/min_terminated_length": 1553.75,
+      "entropy": 0.2867649751715362,
+      "epoch": 0.273817925194072,
+      "geo": 0.7172370981425047,
+      "grad_norm": 0.29296875,
+      "learning_rate": 6.155778894472362e-06,
+      "loss": -0.0,
+      "num_tokens": 91660903.0,
+      "point_2d": 0.75,
+      "point_3d": 0.9462890625,
+      "reward": 0.8695597127079964,
+      "step": 776
+    },
+    {
+      "3d_bbox": 0.6247975081205368,
+      "adv": 0.00014501793702947907,
+      "bbox": 0.9793029651045799,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.125,
+      "completions/max_terminated_length": 1592.125,
+      "completions/mean_length": 1571.03515625,
+      "completions/mean_terminated_length": 1571.03515625,
+      "completions/min_length": 1550.875,
+      "completions/min_terminated_length": 1550.875,
+      "entropy": 0.2792881946079433,
+      "epoch": 0.2766407904022583,
+      "geo": 0.8185150865465403,
+      "grad_norm": 0.2890625,
+      "learning_rate": 6.115577889447236e-06,
+      "loss": 0.0,
+      "num_tokens": 92604352.0,
+      "point_2d": 0.8876953125,
+      "point_3d": 0.8994140625,
+      "reward": 0.8478024452924728,
+      "step": 784
+    },
+    {
+      "3d_bbox": 0.7522557079792023,
+      "adv": 0.00011037173771910602,
+      "bbox": 0.8692253082990646,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.625,
+      "completions/max_terminated_length": 1594.625,
+      "completions/mean_length": 1572.27734375,
+      "completions/mean_terminated_length": 1572.27734375,
+      "completions/min_length": 1552.5,
+      "completions/min_terminated_length": 1552.5,
+      "entropy": 0.2874673060141504,
+      "epoch": 0.2794636556104446,
+      "geo": 0.9408122524619102,
+      "grad_norm": 0.31640625,
+      "learning_rate": 6.07537688442211e-06,
+      "loss": 0.0144,
+      "num_tokens": 93548159.0,
+      "point_2d": 0.7587890625,
+      "point_3d": 0.91796875,
+      "reward": 0.8245597034692764,
+      "step": 792
+    },
+    {
+      "3d_bbox": 0.8205922544002533,
+      "adv": 0.0001977139909286052,
+      "bbox": 0.9971112534403801,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1589.25,
+      "completions/max_terminated_length": 1589.25,
+      "completions/mean_length": 1571.43359375,
+      "completions/mean_terminated_length": 1571.43359375,
+      "completions/min_length": 1552.625,
+      "completions/min_terminated_length": 1552.625,
+      "entropy": 0.28211602522060275,
+      "epoch": 0.2822865208186309,
+      "geo": 0.8445243984460831,
+      "grad_norm": 0.3203125,
+      "learning_rate": 6.035175879396985e-06,
+      "loss": 0.0007,
+      "num_tokens": 94491662.0,
+      "point_2d": 0.884765625,
+      "point_3d": 0.9365234375,
+      "reward": 0.9097481444478035,
+      "step": 800
+    }
+  ],
+  "logging_steps": 8,
+  "max_steps": 2000,
+  "num_input_tokens_seen": 94491662,
+  "num_train_epochs": 1,
+  "save_steps": 50,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-800/video_preprocessor_config.json

@@ -0,0 +1,44 @@
+{
+  "crop_size": null,
+  "data_format": "channels_first",
+  "default_to_square": true,
+  "device": null,
+  "do_center_crop": null,
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_pad": null,
+  "do_rescale": true,
+  "do_resize": true,
+  "do_sample_frames": false,
+  "fps": null,
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "input_data_format": null,
+  "max_frames": 768,
+  "max_pixels": 12845056,
+  "merge_size": 2,
+  "min_frames": 4,
+  "min_pixels": 3136,
+  "num_frames": null,
+  "patch_size": 14,
+  "processor_class": "Qwen2_5_VLProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "return_metadata": false,
+  "size": {
+    "longest_edge": 12845056,
+    "shortest_edge": 3136
+  },
+  "size_divisor": null,
+  "temporal_patch_size": 2,
+  "video_metadata": null,
+  "video_processor_type": "Qwen2VLVideoProcessor"
+}

checkpoint-850/added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

checkpoint-850/chat_template.jinja

@@ -0,0 +1,7 @@
+{% set image_count = namespace(value=0) %}{% set video_count = namespace(value=0) %}{% for message in messages %}{% if loop.first and message['role'] != 'system' %}<|im_start|>system
+You are a helpful assistant.<|im_end|>
+{% endif %}<|im_start|>{{ message['role'] }}
+{% if message['content'] is string %}{{ message['content'] }}<|im_end|>
+{% else %}{% for content in message['content'] %}{% if content['type'] == 'image' or 'image' in content or 'image_url' in content %}{% set image_count.value = image_count.value + 1 %}{% if add_vision_id %}Picture {{ image_count.value }}: {% endif %}<|vision_start|><|image_pad|><|vision_end|>{% elif content['type'] == 'video' or 'video' in content %}{% set video_count.value = video_count.value + 1 %}{% if add_vision_id %}Video {{ video_count.value }}: {% endif %}<|vision_start|><|video_pad|><|vision_end|>{% elif 'text' in content %}{{ content['text'] }}{% endif %}{% endfor %}<|im_end|>
+{% endif %}{% endfor %}{% if add_generation_prompt %}<|im_start|>assistant
+{% endif %}

checkpoint-850/config.json

@@ -0,0 +1,143 @@
+{
+  "architectures": [
+    "Qwen2_5_PVLForConditionalGeneration"
+  ],
+  "attention_dropout": 0.0,
+  "dtype": "bfloat16",
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "image_token_id": 151655,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "max_position_embeddings": 128000,
+  "max_window_layers": 70,
+  "model_type": "qwen2_5_vl",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 36,
+  "num_key_value_heads": 2,
+  "pad_token_id": 151643,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": {
+    "mrope_section": [
+      16,
+      24,
+      24
+    ],
+    "rope_type": "default",
+    "type": "default"
+  },
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "text_config": {
+    "architectures": [
+      "Qwen2_5_VLForConditionalGeneration"
+    ],
+    "attention_dropout": 0.0,
+    "bos_token_id": 151643,
+    "dtype": "bfloat16",
+    "eos_token_id": 151645,
+    "hidden_act": "silu",
+    "hidden_size": 2048,
+    "image_token_id": null,
+    "initializer_range": 0.02,
+    "intermediate_size": 11008,
+    "layer_types": [
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention"
+    ],
+    "max_position_embeddings": 128000,
+    "max_window_layers": 70,
+    "model_type": "qwen2_5_vl_text",
+    "num_attention_heads": 16,
+    "num_hidden_layers": 36,
+    "num_key_value_heads": 2,
+    "rms_norm_eps": 1e-06,
+    "rope_scaling": {
+      "mrope_section": [
+        16,
+        24,
+        24
+      ],
+      "rope_type": "default",
+      "type": "default"
+    },
+    "rope_theta": 1000000.0,
+    "sliding_window": null,
+    "tie_word_embeddings": true,
+    "use_cache": true,
+    "use_sliding_window": false,
+    "video_token_id": null,
+    "vision_end_token_id": 151653,
+    "vision_start_token_id": 151652,
+    "vision_token_id": 151654,
+    "vocab_size": 151936
+  },
+  "transformers_version": "4.56.2",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "video_token_id": 151656,
+  "vision_config": {
+    "depth": 32,
+    "dtype": "bfloat16",
+    "fullatt_block_indexes": [
+      7,
+      15,
+      23,
+      31
+    ],
+    "hidden_act": "silu",
+    "hidden_size": 1280,
+    "in_channels": 3,
+    "in_chans": 3,
+    "initializer_range": 0.02,
+    "intermediate_size": 3420,
+    "model_type": "qwen2_5_vl",
+    "num_heads": 16,
+    "out_hidden_size": 2048,
+    "patch_size": 14,
+    "spatial_merge_size": 2,
+    "spatial_patch_size": 14,
+    "temporal_patch_size": 2,
+    "tokens_per_second": 2,
+    "window_size": 112
+  },
+  "vision_end_token_id": 151653,
+  "vision_start_token_id": 151652,
+  "vision_token_id": 151654,
+  "vocab_size": 151936
+}

checkpoint-850/generation_config.json

@@ -0,0 +1,11 @@
+{
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.05,
+  "temperature": 1e-06,
+  "transformers_version": "4.56.2"
+}

checkpoint-850/preprocessor_config.json

@@ -0,0 +1,37 @@
+{
+  "crop_size": null,
+  "data_format": "channels_first",
+  "default_to_square": true,
+  "device": null,
+  "disable_grouping": null,
+  "do_center_crop": null,
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_processor_type": "Qwen2VLImageProcessorFast",
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "input_data_format": null,
+  "max_pixels": 12845056,
+  "merge_size": 2,
+  "min_pixels": 3136,
+  "patch_size": 14,
+  "processor_class": "Qwen2_5_VLProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "return_tensors": null,
+  "size": {
+    "longest_edge": 12845056,
+    "shortest_edge": 3136
+  },
+  "temporal_patch_size": 2
+}

checkpoint-850/special_tokens_map.json

@@ -0,0 +1,31 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-850/tokenizer_config.json

@@ -0,0 +1,212 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "max_length": null,
+  "model_max_length": 131072,
+  "pad_to_multiple_of": null,
+  "pad_token": "<|endoftext|>",
+  "pad_token_type_id": 0,
+  "padding_side": "left",
+  "processor_class": "Qwen2_5_VLProcessor",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

checkpoint-850/trainer_state.json

@@ -0,0 +1,2472 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.29992942836979536,
+  "eval_steps": 500,
+  "global_step": 850,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "3d_bbox": 0.6610370799899101,
+      "adv": 0.000926484264937244,
+      "bbox": 0.8075725510716438,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.875,
+      "completions/max_terminated_length": 1592.875,
+      "completions/mean_length": 1573.1640625,
+      "completions/mean_terminated_length": 1573.1640625,
+      "completions/min_length": 1554.875,
+      "completions/min_terminated_length": 1554.875,
+      "entropy": 0.2564338077791035,
+      "epoch": 0.0028228652081863093,
+      "geo": 0.7611365988850594,
+      "grad_norm": 0.439453125,
+      "learning_rate": 7e-06,
+      "loss": -0.0002,
+      "num_tokens": 943810.0,
+      "point_2d": 0.875,
+      "point_3d": 0.638671875,
+      "reward": 0.7455703690648079,
+      "step": 8
+    },
+    {
+      "3d_bbox": 0.6985021606087685,
+      "adv": 0.0008333708265126916,
+      "bbox": 0.9951911941170692,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.25,
+      "completions/max_terminated_length": 1594.25,
+      "completions/mean_length": 1573.64453125,
+      "completions/mean_terminated_length": 1573.64453125,
+      "completions/min_length": 1552.375,
+      "completions/min_terminated_length": 1552.375,
+      "entropy": 0.2595143502112478,
+      "epoch": 0.0056457304163726185,
+      "geo": 0.8901385143399239,
+      "grad_norm": 0.341796875,
+      "learning_rate": 9.974874371859297e-06,
+      "loss": 0.0564,
+      "num_tokens": 1888015.0,
+      "point_2d": 1.0,
+      "point_3d": 0.5947265625,
+      "reward": 0.8221049904823303,
+      "step": 16
+    },
+    {
+      "3d_bbox": 0.6706115230917931,
+      "adv": -0.11667590066201683,
+      "bbox": 0.9641365110874176,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1588.75,
+      "completions/max_terminated_length": 1588.75,
+      "completions/mean_length": 1568.69921875,
+      "completions/mean_terminated_length": 1568.69921875,
+      "completions/min_length": 1549.625,
+      "completions/min_terminated_length": 1549.625,
+      "entropy": 0.2620172018650919,
+      "epoch": 0.008468595624558928,
+      "geo": 0.8719085603952408,
+      "grad_norm": 0.31640625,
+      "learning_rate": 9.934673366834172e-06,
+      "loss": 0.0434,
+      "num_tokens": 2830754.0,
+      "point_2d": 0.96484375,
+      "point_3d": 0.7060546875,
+      "reward": 0.8264116197824478,
+      "step": 24
+    },
+    {
+      "3d_bbox": 0.6768141128122807,
+      "adv": 0.00011848201484099263,
+      "bbox": 0.8640391379594803,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.875,
+      "completions/max_terminated_length": 1599.875,
+      "completions/mean_length": 1579.40625,
+      "completions/mean_terminated_length": 1579.40625,
+      "completions/min_length": 1563.0,
+      "completions/min_terminated_length": 1563.0,
+      "entropy": 0.26498199813067913,
+      "epoch": 0.011291460832745237,
+      "geo": 0.8661364912986755,
+      "grad_norm": 0.36328125,
+      "learning_rate": 9.894472361809046e-06,
+      "loss": -0.0005,
+      "num_tokens": 3776674.0,
+      "point_2d": 0.8642578125,
+      "point_3d": 0.603515625,
+      "reward": 0.752156674861908,
+      "step": 32
+    },
+    {
+      "3d_bbox": 0.6714089289307594,
+      "adv": -0.000722882674608627,
+      "bbox": 0.9624817371368408,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1593.75,
+      "completions/max_terminated_length": 1593.75,
+      "completions/mean_length": 1575.52734375,
+      "completions/mean_terminated_length": 1575.52734375,
+      "completions/min_length": 1559.875,
+      "completions/min_terminated_length": 1559.875,
+      "entropy": 0.2738626003265381,
+      "epoch": 0.014114326040931546,
+      "geo": 0.8932898789644241,
+      "grad_norm": 0.341796875,
+      "learning_rate": 9.85427135678392e-06,
+      "loss": 0.0002,
+      "num_tokens": 4721169.0,
+      "point_2d": 0.99609375,
+      "point_3d": 0.6064453125,
+      "reward": 0.8091074377298355,
+      "step": 40
+    },
+    {
+      "3d_bbox": 0.6575617864727974,
+      "adv": -0.0007460190936399158,
+      "bbox": 0.9782034233212471,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.75,
+      "completions/max_terminated_length": 1598.75,
+      "completions/mean_length": 1581.515625,
+      "completions/mean_terminated_length": 1581.515625,
+      "completions/min_length": 1563.875,
+      "completions/min_terminated_length": 1563.875,
+      "entropy": 0.26874606031924486,
+      "epoch": 0.016937191249117856,
+      "geo": 0.922828696668148,
+      "grad_norm": 0.384765625,
+      "learning_rate": 9.814070351758794e-06,
+      "loss": 0.0001,
+      "num_tokens": 5667501.0,
+      "point_2d": 0.9404296875,
+      "point_3d": 0.6171875,
+      "reward": 0.7983455955982208,
+      "step": 48
+    },
+    {
+      "3d_bbox": 0.7516707628965378,
+      "adv": 0.0002944665375252953,
+      "bbox": 0.870270783547312,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1605.5,
+      "completions/max_terminated_length": 1605.5,
+      "completions/mean_length": 1583.109375,
+      "completions/mean_terminated_length": 1583.109375,
+      "completions/min_length": 1563.125,
+      "completions/min_terminated_length": 1563.125,
+      "entropy": 0.27441642340272665,
+      "epoch": 0.019760056457304165,
+      "geo": 0.8707154765725136,
+      "grad_norm": 0.35546875,
+      "learning_rate": 9.773869346733669e-06,
+      "loss": 0.0002,
+      "num_tokens": 6614089.0,
+      "point_2d": 0.875,
+      "point_3d": 0.728515625,
+      "reward": 0.8063642866909504,
+      "step": 56
+    },
+    {
+      "3d_bbox": 0.7704295367002487,
+      "adv": 0.0006564971263287589,
+      "bbox": 0.8711516009643674,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1604.75,
+      "completions/max_terminated_length": 1604.75,
+      "completions/mean_length": 1582.09375,
+      "completions/mean_terminated_length": 1582.09375,
+      "completions/min_length": 1562.25,
+      "completions/min_terminated_length": 1562.25,
+      "entropy": 0.2721484359353781,
+      "epoch": 0.022582921665490474,
+      "geo": 0.8998556286096573,
+      "grad_norm": 0.373046875,
+      "learning_rate": 9.733668341708543e-06,
+      "loss": 0.0156,
+      "num_tokens": 7560513.0,
+      "point_2d": 0.875,
+      "point_3d": 0.7880859375,
+      "reward": 0.8261667564511299,
+      "step": 64
+    },
+    {
+      "3d_bbox": 0.6965150758624077,
+      "adv": 5.7399442994210403e-05,
+      "bbox": 0.9924400746822357,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1603.875,
+      "completions/max_terminated_length": 1603.875,
+      "completions/mean_length": 1580.8203125,
+      "completions/mean_terminated_length": 1580.8203125,
+      "completions/min_length": 1562.0,
+      "completions/min_terminated_length": 1562.0,
+      "entropy": 0.2694994225166738,
+      "epoch": 0.025405786873676783,
+      "geo": 0.7992948293685913,
+      "grad_norm": 0.40625,
+      "learning_rate": 9.693467336683417e-06,
+      "loss": -0.0004,
+      "num_tokens": 8506787.0,
+      "point_2d": 0.875,
+      "point_3d": 0.6201171875,
+      "reward": 0.796018086373806,
+      "step": 72
+    },
+    {
+      "3d_bbox": 0.9022366181015968,
+      "adv": 0.000732647989934776,
+      "bbox": 0.9670402258634567,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.5,
+      "completions/max_terminated_length": 1598.5,
+      "completions/mean_length": 1577.83203125,
+      "completions/mean_terminated_length": 1577.83203125,
+      "completions/min_length": 1560.0,
+      "completions/min_terminated_length": 1560.0,
+      "entropy": 0.2729479926638305,
+      "epoch": 0.028228652081863093,
+      "geo": 0.770210200920701,
+      "grad_norm": 0.28515625,
+      "learning_rate": 9.653266331658291e-06,
+      "loss": 0.0285,
+      "num_tokens": 9452032.0,
+      "point_2d": 0.7578125,
+      "point_3d": 0.8740234375,
+      "reward": 0.8752781972289085,
+      "step": 80
+    },
+    {
+      "3d_bbox": 0.8243749290704727,
+      "adv": 0.0004942354327113208,
+      "bbox": 0.9483200311660767,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.0,
+      "completions/max_terminated_length": 1599.0,
+      "completions/mean_length": 1578.5703125,
+      "completions/mean_terminated_length": 1578.5703125,
+      "completions/min_length": 1564.875,
+      "completions/min_terminated_length": 1564.875,
+      "entropy": 0.27196697471663356,
+      "epoch": 0.031051517290049402,
+      "geo": 0.8520620986819267,
+      "grad_norm": 0.337890625,
+      "learning_rate": 9.613065326633166e-06,
+      "loss": 0.0,
+      "num_tokens": 10397602.0,
+      "point_2d": 0.84375,
+      "point_3d": 0.7431640625,
+      "reward": 0.8399022594094276,
+      "step": 88
+    },
+    {
+      "3d_bbox": 0.7930564805865288,
+      "adv": -0.05789744130743202,
+      "bbox": 0.9452124014496803,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1600.125,
+      "completions/max_terminated_length": 1600.125,
+      "completions/mean_length": 1579.25390625,
+      "completions/mean_terminated_length": 1579.25390625,
+      "completions/min_length": 1560.25,
+      "completions/min_terminated_length": 1560.25,
+      "entropy": 0.2732193539850414,
+      "epoch": 0.03387438249823571,
+      "geo": 0.8466203399002552,
+      "grad_norm": 0.35546875,
+      "learning_rate": 9.57286432160804e-06,
+      "loss": 0.0143,
+      "num_tokens": 11343267.0,
+      "point_2d": 0.7138671875,
+      "point_3d": 0.8203125,
+      "reward": 0.8181121572852135,
+      "step": 96
+    },
+    {
+      "3d_bbox": 0.6107643470168114,
+      "adv": -0.0019174332555849105,
+      "bbox": 0.9693493247032166,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1605.375,
+      "completions/max_terminated_length": 1605.375,
+      "completions/mean_length": 1580.6875,
+      "completions/mean_terminated_length": 1580.6875,
+      "completions/min_length": 1559.75,
+      "completions/min_terminated_length": 1559.75,
+      "entropy": 0.2712484383955598,
+      "epoch": 0.03669724770642202,
+      "geo": 0.760684534907341,
+      "grad_norm": 0.392578125,
+      "learning_rate": 9.532663316582916e-06,
+      "loss": 0.0006,
+      "num_tokens": 12289259.0,
+      "point_2d": 0.849609375,
+      "point_3d": 0.7109375,
+      "reward": 0.7851651459932327,
+      "step": 104
+    },
+    {
+      "3d_bbox": 0.7011105790734291,
+      "adv": 0.000332610031364311,
+      "bbox": 0.997136726975441,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.5,
+      "completions/max_terminated_length": 1601.5,
+      "completions/mean_length": 1578.38671875,
+      "completions/mean_terminated_length": 1578.38671875,
+      "completions/min_length": 1560.0,
+      "completions/min_terminated_length": 1560.0,
+      "entropy": 0.2741631781682372,
+      "epoch": 0.03952011291460833,
+      "geo": 0.868082121014595,
+      "grad_norm": 0.349609375,
+      "learning_rate": 9.49246231155779e-06,
+      "loss": -0.0005,
+      "num_tokens": 13234494.0,
+      "point_2d": 0.875,
+      "point_3d": 0.759765625,
+      "reward": 0.8332532122731209,
+      "step": 112
+    },
+    {
+      "3d_bbox": 0.6414853371679783,
+      "adv": 6.49127487122314e-05,
+      "bbox": 0.9946233779191971,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.875,
+      "completions/max_terminated_length": 1599.875,
+      "completions/mean_length": 1580.48828125,
+      "completions/mean_terminated_length": 1580.48828125,
+      "completions/min_length": 1562.5,
+      "completions/min_terminated_length": 1562.5,
+      "entropy": 0.2765682805329561,
+      "epoch": 0.04234297812279464,
+      "geo": 0.7812720276415348,
+      "grad_norm": 0.392578125,
+      "learning_rate": 9.452261306532664e-06,
+      "loss": -0.0002,
+      "num_tokens": 14180739.0,
+      "point_2d": 0.875,
+      "point_3d": 0.7275390625,
+      "reward": 0.8096619322896004,
+      "step": 120
+    },
+    {
+      "3d_bbox": 0.696478821337223,
+      "adv": 0.00018893864194069465,
+      "bbox": 0.9915411844849586,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1600.75,
+      "completions/max_terminated_length": 1600.75,
+      "completions/mean_length": 1581.48828125,
+      "completions/mean_terminated_length": 1581.48828125,
+      "completions/min_length": 1566.25,
+      "completions/min_terminated_length": 1566.25,
+      "entropy": 0.27939224475994706,
+      "epoch": 0.04516584333098095,
+      "geo": 0.9225678443908691,
+      "grad_norm": 0.359375,
+      "learning_rate": 9.412060301507539e-06,
+      "loss": 0.0004,
+      "num_tokens": 15126808.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8037109375,
+      "reward": 0.8729327321052551,
+      "step": 128
+    },
+    {
+      "3d_bbox": 0.7811713069677353,
+      "adv": -0.00017097587488024146,
+      "bbox": 0.983598880469799,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.0,
+      "completions/max_terminated_length": 1599.0,
+      "completions/mean_length": 1575.9765625,
+      "completions/mean_terminated_length": 1575.9765625,
+      "completions/min_length": 1560.875,
+      "completions/min_terminated_length": 1560.875,
+      "entropy": 0.2712345658801496,
+      "epoch": 0.04798870853916726,
+      "geo": 0.829373586922884,
+      "grad_norm": 0.365234375,
+      "learning_rate": 9.371859296482413e-06,
+      "loss": 0.0002,
+      "num_tokens": 16071682.0,
+      "point_2d": 0.75,
+      "point_3d": 0.76171875,
+      "reward": 0.819122239947319,
+      "step": 136
+    },
+    {
+      "3d_bbox": 0.6566238105297089,
+      "adv": 5.832723485355018e-05,
+      "bbox": 0.9859664067626,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1597.5,
+      "completions/max_terminated_length": 1597.5,
+      "completions/mean_length": 1580.12890625,
+      "completions/mean_terminated_length": 1580.12890625,
+      "completions/min_length": 1565.0,
+      "completions/min_terminated_length": 1565.0,
+      "entropy": 0.2754409732297063,
+      "epoch": 0.05081157374735357,
+      "geo": 0.8895541727542877,
+      "grad_norm": 0.318359375,
+      "learning_rate": 9.331658291457287e-06,
+      "loss": 0.0218,
+      "num_tokens": 17017499.0,
+      "point_2d": 1.0,
+      "point_3d": 0.6328125,
+      "reward": 0.8188506811857224,
+      "step": 144
+    },
+    {
+      "3d_bbox": 0.655942264944315,
+      "adv": 0.00024080171715468168,
+      "bbox": 0.9830196872353554,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1600.125,
+      "completions/max_terminated_length": 1600.125,
+      "completions/mean_length": 1581.3828125,
+      "completions/mean_terminated_length": 1581.3828125,
+      "completions/min_length": 1564.0,
+      "completions/min_terminated_length": 1564.0,
+      "entropy": 0.2764031286351383,
+      "epoch": 0.053634438955539876,
+      "geo": 0.8772727139294147,
+      "grad_norm": 0.294921875,
+      "learning_rate": 9.291457286432161e-06,
+      "loss": 0.015,
+      "num_tokens": 17963757.0,
+      "point_2d": 0.875,
+      "point_3d": 0.7880859375,
+      "reward": 0.8255119547247887,
+      "step": 152
+    },
+    {
+      "3d_bbox": 0.8087940737605095,
+      "adv": -0.062093503400319605,
+      "bbox": 0.9799230694770813,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.5,
+      "completions/max_terminated_length": 1598.5,
+      "completions/mean_length": 1579.2421875,
+      "completions/mean_terminated_length": 1579.2421875,
+      "completions/min_length": 1560.25,
+      "completions/min_terminated_length": 1560.25,
+      "entropy": 0.27893486013635993,
+      "epoch": 0.056457304163726185,
+      "geo": 0.8533996716141701,
+      "grad_norm": 0.314453125,
+      "learning_rate": 9.251256281407036e-06,
+      "loss": 0.0157,
+      "num_tokens": 18909307.0,
+      "point_2d": 0.859375,
+      "point_3d": 0.814453125,
+      "reward": 0.8656363189220428,
+      "step": 160
+    },
+    {
+      "3d_bbox": 0.7417622283101082,
+      "adv": 0.00038278038937278325,
+      "bbox": 0.9947961643338203,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.875,
+      "completions/max_terminated_length": 1596.875,
+      "completions/mean_length": 1578.4765625,
+      "completions/mean_terminated_length": 1578.4765625,
+      "completions/min_length": 1562.375,
+      "completions/min_terminated_length": 1562.375,
+      "entropy": 0.2772809020243585,
+      "epoch": 0.059280169371912494,
+      "geo": 0.8446554020047188,
+      "grad_norm": 0.337890625,
+      "learning_rate": 9.21105527638191e-06,
+      "loss": -0.0001,
+      "num_tokens": 19854893.0,
+      "point_2d": 0.875,
+      "point_3d": 0.7666015625,
+      "reward": 0.844539999961853,
+      "step": 168
+    },
+    {
+      "3d_bbox": 0.7199448123574257,
+      "adv": -0.00025279237161157653,
+      "bbox": 0.9939833134412766,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.375,
+      "completions/max_terminated_length": 1599.375,
+      "completions/mean_length": 1577.875,
+      "completions/mean_terminated_length": 1577.875,
+      "completions/min_length": 1560.875,
+      "completions/min_terminated_length": 1560.875,
+      "entropy": 0.2748410953208804,
+      "epoch": 0.062103034580098804,
+      "geo": 0.8426482602953911,
+      "grad_norm": 0.353515625,
+      "learning_rate": 9.170854271356784e-06,
+      "loss": 0.0001,
+      "num_tokens": 20800093.0,
+      "point_2d": 0.8935546875,
+      "point_3d": 0.7666015625,
+      "reward": 0.8435210958123207,
+      "step": 176
+    },
+    {
+      "3d_bbox": 0.8168098330497742,
+      "adv": 0.0003567840885807527,
+      "bbox": 0.8971930183470249,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.5,
+      "completions/max_terminated_length": 1601.5,
+      "completions/mean_length": 1581.01171875,
+      "completions/mean_terminated_length": 1581.01171875,
+      "completions/min_length": 1562.625,
+      "completions/min_terminated_length": 1562.625,
+      "entropy": 0.27716846112161875,
+      "epoch": 0.06492589978828511,
+      "geo": 0.8458116948604584,
+      "grad_norm": 0.376953125,
+      "learning_rate": 9.13065326633166e-06,
+      "loss": -0.0,
+      "num_tokens": 21746232.0,
+      "point_2d": 0.736328125,
+      "point_3d": 0.859375,
+      "reward": 0.8274264857172966,
+      "step": 184
+    },
+    {
+      "3d_bbox": 0.7658436857163906,
+      "adv": -0.00043575108065851964,
+      "bbox": 0.994478665292263,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.5,
+      "completions/max_terminated_length": 1599.5,
+      "completions/mean_length": 1577.2265625,
+      "completions/mean_terminated_length": 1577.2265625,
+      "completions/min_length": 1558.25,
+      "completions/min_terminated_length": 1558.25,
+      "entropy": 0.28411462996155024,
+      "epoch": 0.06774876499647142,
+      "geo": 0.8858354687690735,
+      "grad_norm": 0.404296875,
+      "learning_rate": 9.090452261306534e-06,
+      "loss": -0.0,
+      "num_tokens": 22691162.0,
+      "point_2d": 1.0,
+      "point_3d": 0.83203125,
+      "reward": 0.8980883881449699,
+      "step": 192
+    },
+    {
+      "3d_bbox": 0.689879797399044,
+      "adv": 0.0008139211470279406,
+      "bbox": 0.8756355233490467,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.25,
+      "completions/max_terminated_length": 1595.25,
+      "completions/mean_length": 1579.328125,
+      "completions/mean_terminated_length": 1579.328125,
+      "completions/min_length": 1562.125,
+      "completions/min_terminated_length": 1562.125,
+      "entropy": 0.27792059257626534,
+      "epoch": 0.07057163020465773,
+      "geo": 0.8776071295142174,
+      "grad_norm": 0.28125,
+      "learning_rate": 9.050251256281409e-06,
+      "loss": -0.0006,
+      "num_tokens": 23637118.0,
+      "point_2d": 0.875,
+      "point_3d": 0.765625,
+      "reward": 0.8015350699424744,
+      "step": 200
+    },
+    {
+      "3d_bbox": 0.7628588750958443,
+      "adv": 0.0005178588535272866,
+      "bbox": 0.9862369671463966,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1603.5,
+      "completions/max_terminated_length": 1603.5,
+      "completions/mean_length": 1582.5,
+      "completions/mean_terminated_length": 1582.5,
+      "completions/min_length": 1561.75,
+      "completions/min_terminated_length": 1561.75,
+      "entropy": 0.2777437614277005,
+      "epoch": 0.07339449541284404,
+      "geo": 0.7896297052502632,
+      "grad_norm": 0.318359375,
+      "learning_rate": 9.010050251256283e-06,
+      "loss": -0.0003,
+      "num_tokens": 24583790.0,
+      "point_2d": 0.75,
+      "point_3d": 0.822265625,
+      "reward": 0.8303403630852699,
+      "step": 208
+    },
+    {
+      "3d_bbox": 0.6901299897581339,
+      "adv": 1.327232166659087e-05,
+      "bbox": 0.9742533341050148,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.375,
+      "completions/max_terminated_length": 1595.375,
+      "completions/mean_length": 1580.3828125,
+      "completions/mean_terminated_length": 1580.3828125,
+      "completions/min_length": 1566.125,
+      "completions/min_terminated_length": 1566.125,
+      "entropy": 0.2789337821304798,
+      "epoch": 0.07621736062103035,
+      "geo": 0.9093400537967682,
+      "grad_norm": 0.33984375,
+      "learning_rate": 8.969849246231157e-06,
+      "loss": 0.0001,
+      "num_tokens": 25529904.0,
+      "point_2d": 1.0,
+      "point_3d": 0.724609375,
+      "reward": 0.8472481593489647,
+      "step": 216
+    },
+    {
+      "3d_bbox": 0.8317120671272278,
+      "adv": 0.0007813970478309784,
+      "bbox": 0.99237409979105,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.375,
+      "completions/max_terminated_length": 1592.375,
+      "completions/mean_length": 1576.484375,
+      "completions/mean_terminated_length": 1576.484375,
+      "completions/min_length": 1559.125,
+      "completions/min_terminated_length": 1559.125,
+      "entropy": 0.281503742095083,
+      "epoch": 0.07904022582921666,
+      "geo": 0.9148535653948784,
+      "grad_norm": 0.33203125,
+      "learning_rate": 8.92964824120603e-06,
+      "loss": -0.0005,
+      "num_tokens": 26474884.0,
+      "point_2d": 1.0,
+      "point_3d": 0.859375,
+      "reward": 0.9208653047680855,
+      "step": 224
+    },
+    {
+      "3d_bbox": 0.8539343103766441,
+      "adv": 0.0006083497164581786,
+      "bbox": 0.9947174340486526,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1607.25,
+      "completions/max_terminated_length": 1607.25,
+      "completions/mean_length": 1582.21875,
+      "completions/mean_terminated_length": 1582.21875,
+      "completions/min_length": 1562.75,
+      "completions/min_terminated_length": 1562.75,
+      "entropy": 0.28075447864830494,
+      "epoch": 0.08186309103740297,
+      "geo": 0.9205377101898193,
+      "grad_norm": 0.3828125,
+      "learning_rate": 8.889447236180904e-06,
+      "loss": 0.0151,
+      "num_tokens": 27421356.0,
+      "point_2d": 0.9423828125,
+      "point_3d": 0.919921875,
+      "reward": 0.9277391210198402,
+      "step": 232
+    },
+    {
+      "3d_bbox": 0.8143627047538757,
+      "adv": 0.0002584877729532309,
+      "bbox": 0.9961168393492699,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1604.375,
+      "completions/max_terminated_length": 1604.375,
+      "completions/mean_length": 1581.84375,
+      "completions/mean_terminated_length": 1581.84375,
+      "completions/min_length": 1562.0,
+      "completions/min_terminated_length": 1562.0,
+      "entropy": 0.2808487247675657,
+      "epoch": 0.08468595624558928,
+      "geo": 0.9482192173600197,
+      "grad_norm": 0.3125,
+      "learning_rate": 8.84924623115578e-06,
+      "loss": 0.015,
+      "num_tokens": 28367588.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8505859375,
+      "reward": 0.9152663722634315,
+      "step": 240
+    },
+    {
+      "3d_bbox": 0.7562066540122032,
+      "adv": 0.0011895421193912625,
+      "bbox": 0.9870459213852882,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1605.625,
+      "completions/max_terminated_length": 1605.625,
+      "completions/mean_length": 1582.22265625,
+      "completions/mean_terminated_length": 1582.22265625,
+      "completions/min_length": 1560.125,
+      "completions/min_terminated_length": 1560.125,
+      "entropy": 0.2828774293884635,
+      "epoch": 0.08750882145377559,
+      "geo": 0.8906307443976402,
+      "grad_norm": 0.310546875,
+      "learning_rate": 8.809045226130654e-06,
+      "loss": -0.0001,
+      "num_tokens": 29314053.0,
+      "point_2d": 1.0,
+      "point_3d": 0.84765625,
+      "reward": 0.8977271988987923,
+      "step": 248
+    },
+    {
+      "3d_bbox": 0.662892609834671,
+      "adv": -0.0004179688154977157,
+      "bbox": 0.9923923686146736,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.875,
+      "completions/max_terminated_length": 1598.875,
+      "completions/mean_length": 1579.796875,
+      "completions/mean_terminated_length": 1579.796875,
+      "completions/min_length": 1563.125,
+      "completions/min_terminated_length": 1563.125,
+      "entropy": 0.27959624165669084,
+      "epoch": 0.0903316866619619,
+      "geo": 0.9055434167385101,
+      "grad_norm": 0.259765625,
+      "learning_rate": 8.768844221105528e-06,
+      "loss": -0.0002,
+      "num_tokens": 30259921.0,
+      "point_2d": 0.7509765625,
+      "point_3d": 0.8154296875,
+      "reward": 0.805422805249691,
+      "step": 256
+    },
+    {
+      "3d_bbox": 0.7727803811430931,
+      "adv": -0.0010380800276834634,
+      "bbox": 0.9807735681533813,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1604.375,
+      "completions/max_terminated_length": 1604.375,
+      "completions/mean_length": 1581.1796875,
+      "completions/mean_terminated_length": 1581.1796875,
+      "completions/min_length": 1564.375,
+      "completions/min_terminated_length": 1564.375,
+      "entropy": 0.28243178501725197,
+      "epoch": 0.0931545518701482,
+      "geo": 0.783092018100433,
+      "grad_norm": 0.3203125,
+      "learning_rate": 8.728643216080403e-06,
+      "loss": 0.0003,
+      "num_tokens": 31205887.0,
+      "point_2d": 0.75,
+      "point_3d": 0.8544921875,
+      "reward": 0.8395115286111832,
+      "step": 264
+    },
+    {
+      "3d_bbox": 0.7379852011799812,
+      "adv": -0.0005326499958755448,
+      "bbox": 0.9934581220149994,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.875,
+      "completions/max_terminated_length": 1598.875,
+      "completions/mean_length": 1579.44921875,
+      "completions/mean_terminated_length": 1579.44921875,
+      "completions/min_length": 1560.375,
+      "completions/min_terminated_length": 1560.375,
+      "entropy": 0.2767978576011956,
+      "epoch": 0.09597741707833451,
+      "geo": 0.9199443385004997,
+      "grad_norm": 0.40625,
+      "learning_rate": 8.688442211055277e-06,
+      "loss": 0.0002,
+      "num_tokens": 32151650.0,
+      "point_2d": 0.9990234375,
+      "point_3d": 0.8896484375,
+      "reward": 0.9050287976861,
+      "step": 272
+    },
+    {
+      "3d_bbox": 0.6362542062997818,
+      "adv": 0.0003683161558001302,
+      "bbox": 0.9273895137012005,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1603.375,
+      "completions/max_terminated_length": 1603.375,
+      "completions/mean_length": 1583.1875,
+      "completions/mean_terminated_length": 1583.1875,
+      "completions/min_length": 1566.25,
+      "completions/min_terminated_length": 1566.25,
+      "entropy": 0.28121166303753853,
+      "epoch": 0.09880028228652082,
+      "geo": 0.6698016822338104,
+      "grad_norm": 0.419921875,
+      "learning_rate": 8.648241206030151e-06,
+      "loss": 0.0104,
+      "num_tokens": 33098514.0,
+      "point_2d": 0.748046875,
+      "point_3d": 0.8564453125,
+      "reward": 0.7920339703559875,
+      "step": 280
+    },
+    {
+      "3d_bbox": 0.8045329824090004,
+      "adv": 0.0005392557122831931,
+      "bbox": 0.9957676902413368,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.75,
+      "completions/max_terminated_length": 1598.75,
+      "completions/mean_length": 1582.8515625,
+      "completions/mean_terminated_length": 1582.8515625,
+      "completions/min_length": 1564.75,
+      "completions/min_terminated_length": 1564.75,
+      "entropy": 0.2840573275461793,
+      "epoch": 0.10162314749470713,
+      "geo": 0.894830085337162,
+      "grad_norm": 0.263671875,
+      "learning_rate": 8.608040201005025e-06,
+      "loss": -0.0003,
+      "num_tokens": 34045156.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8974609375,
+      "reward": 0.9244403913617134,
+      "step": 288
+    },
+    {
+      "3d_bbox": 0.6487490832805634,
+      "adv": -0.036328162997961044,
+      "bbox": 0.9537674784660339,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1606.125,
+      "completions/max_terminated_length": 1606.125,
+      "completions/mean_length": 1583.43359375,
+      "completions/mean_terminated_length": 1583.43359375,
+      "completions/min_length": 1564.25,
+      "completions/min_terminated_length": 1564.25,
+      "entropy": 0.2789871282875538,
+      "epoch": 0.10444601270289344,
+      "geo": 0.8219972774386406,
+      "grad_norm": 0.3671875,
+      "learning_rate": 8.5678391959799e-06,
+      "loss": 0.0091,
+      "num_tokens": 34991979.0,
+      "point_2d": 0.875,
+      "point_3d": 0.8125,
+      "reward": 0.8225041329860687,
+      "step": 296
+    },
+    {
+      "3d_bbox": 0.7852863818407059,
+      "adv": -2.922765997936949e-05,
+      "bbox": 0.9355472847819328,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.5,
+      "completions/max_terminated_length": 1601.5,
+      "completions/mean_length": 1579.7890625,
+      "completions/mean_terminated_length": 1579.7890625,
+      "completions/min_length": 1561.5,
+      "completions/min_terminated_length": 1561.5,
+      "entropy": 0.2808959139510989,
+      "epoch": 0.10726887791107975,
+      "geo": 0.7100357487797737,
+      "grad_norm": 0.333984375,
+      "learning_rate": 8.527638190954774e-06,
+      "loss": 0.0003,
+      "num_tokens": 35937853.0,
+      "point_2d": 0.6728515625,
+      "point_3d": 0.798828125,
+      "reward": 0.7981283441185951,
+      "step": 304
+    },
+    {
+      "3d_bbox": 0.6897341758012772,
+      "adv": -0.0010775219143397408,
+      "bbox": 0.9957167953252792,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.0,
+      "completions/max_terminated_length": 1598.0,
+      "completions/mean_length": 1578.40234375,
+      "completions/mean_terminated_length": 1578.40234375,
+      "completions/min_length": 1560.625,
+      "completions/min_terminated_length": 1560.625,
+      "entropy": 0.28013237891718745,
+      "epoch": 0.11009174311926606,
+      "geo": 0.8601852916181087,
+      "grad_norm": 0.322265625,
+      "learning_rate": 8.487437185929648e-06,
+      "loss": 0.0003,
+      "num_tokens": 36883412.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8154296875,
+      "reward": 0.8752201721072197,
+      "step": 312
+    },
+    {
+      "3d_bbox": 0.7732156217098236,
+      "adv": 0.00015621728016412817,
+      "bbox": 0.9900884553790092,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.25,
+      "completions/max_terminated_length": 1595.25,
+      "completions/mean_length": 1574.06640625,
+      "completions/mean_terminated_length": 1574.06640625,
+      "completions/min_length": 1556.5,
+      "completions/min_terminated_length": 1556.5,
+      "entropy": 0.2842056662775576,
+      "epoch": 0.11291460832745237,
+      "geo": 0.8583031129091978,
+      "grad_norm": 0.359375,
+      "learning_rate": 8.447236180904524e-06,
+      "loss": -0.0003,
+      "num_tokens": 37827701.0,
+      "point_2d": 0.8583984375,
+      "point_3d": 0.8681640625,
+      "reward": 0.8724666386842728,
+      "step": 320
+    },
+    {
+      "3d_bbox": 0.7427574843168259,
+      "adv": -0.00022968700068304315,
+      "bbox": 0.9458357989788055,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.25,
+      "completions/max_terminated_length": 1599.25,
+      "completions/mean_length": 1576.78125,
+      "completions/mean_terminated_length": 1576.78125,
+      "completions/min_length": 1555.625,
+      "completions/min_terminated_length": 1555.625,
+      "entropy": 0.28632913110777736,
+      "epoch": 0.11573747353563868,
+      "geo": 0.9527824521064758,
+      "grad_norm": 0.2578125,
+      "learning_rate": 8.407035175879398e-06,
+      "loss": -0.0,
+      "num_tokens": 38772853.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8623046875,
+      "reward": 0.8877245113253593,
+      "step": 328
+    },
+    {
+      "3d_bbox": 0.7519282810389996,
+      "adv": 0.0003111484527380526,
+      "bbox": 0.9336327686905861,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.0,
+      "completions/max_terminated_length": 1598.0,
+      "completions/mean_length": 1578.54296875,
+      "completions/mean_terminated_length": 1578.54296875,
+      "completions/min_length": 1560.0,
+      "completions/min_terminated_length": 1560.0,
+      "entropy": 0.2819040766917169,
+      "epoch": 0.11856033874382499,
+      "geo": 0.9259950742125511,
+      "grad_norm": 0.306640625,
+      "learning_rate": 8.366834170854273e-06,
+      "loss": 0.0202,
+      "num_tokens": 39718400.0,
+      "point_2d": 0.7451171875,
+      "point_3d": 0.748046875,
+      "reward": 0.7946812808513641,
+      "step": 336
+    },
+    {
+      "3d_bbox": 0.6988782063126564,
+      "adv": 8.153711894226534e-05,
+      "bbox": 0.9918540865182877,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1602.0,
+      "completions/max_terminated_length": 1602.0,
+      "completions/mean_length": 1578.3984375,
+      "completions/mean_terminated_length": 1578.3984375,
+      "completions/min_length": 1557.0,
+      "completions/min_terminated_length": 1557.0,
+      "entropy": 0.2852953290566802,
+      "epoch": 0.1213832039520113,
+      "geo": 0.824953593313694,
+      "grad_norm": 0.33984375,
+      "learning_rate": 8.326633165829147e-06,
+      "loss": 0.0003,
+      "num_tokens": 40663774.0,
+      "point_2d": 0.875,
+      "point_3d": 0.771484375,
+      "reward": 0.8343041799962521,
+      "step": 344
+    },
+    {
+      "3d_bbox": 0.7085578441619873,
+      "adv": 0.000511349891894497,
+      "bbox": 0.9954689294099808,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.625,
+      "completions/max_terminated_length": 1601.625,
+      "completions/mean_length": 1578.21484375,
+      "completions/mean_terminated_length": 1578.21484375,
+      "completions/min_length": 1557.5,
+      "completions/min_terminated_length": 1557.5,
+      "entropy": 0.2835490182042122,
+      "epoch": 0.12420606916019761,
+      "geo": 0.8287867829203606,
+      "grad_norm": 0.31640625,
+      "learning_rate": 8.286432160804021e-06,
+      "loss": 0.0,
+      "num_tokens": 41609261.0,
+      "point_2d": 0.875,
+      "point_3d": 0.8564453125,
+      "reward": 0.8588679879903793,
+      "step": 352
+    },
+    {
+      "3d_bbox": 0.7324622049927711,
+      "adv": -0.06052318853471661,
+      "bbox": 0.9785870239138603,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.75,
+      "completions/max_terminated_length": 1599.75,
+      "completions/mean_length": 1580.8671875,
+      "completions/mean_terminated_length": 1580.8671875,
+      "completions/min_length": 1562.875,
+      "completions/min_terminated_length": 1562.875,
+      "entropy": 0.2788377907127142,
+      "epoch": 0.12702893436838392,
+      "geo": 0.8886202499270439,
+      "grad_norm": 0.765625,
+      "learning_rate": 8.246231155778895e-06,
+      "loss": 0.0295,
+      "num_tokens": 42555355.0,
+      "point_2d": 0.984375,
+      "point_3d": 0.80078125,
+      "reward": 0.8740513622760773,
+      "step": 360
+    },
+    {
+      "3d_bbox": 0.6365351751446724,
+      "adv": 0.0003275554918218404,
+      "bbox": 0.8720735665410757,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.875,
+      "completions/max_terminated_length": 1601.875,
+      "completions/mean_length": 1577.73046875,
+      "completions/mean_terminated_length": 1577.73046875,
+      "completions/min_length": 1555.75,
+      "completions/min_terminated_length": 1555.75,
+      "entropy": 0.28446380188688636,
+      "epoch": 0.12985179957657023,
+      "geo": 0.8144352469826117,
+      "grad_norm": 0.34375,
+      "learning_rate": 8.20603015075377e-06,
+      "loss": -0.0001,
+      "num_tokens": 43500534.0,
+      "point_2d": 0.875,
+      "point_3d": 0.7666015625,
+      "reward": 0.787552572786808,
+      "step": 368
+    },
+    {
+      "3d_bbox": 0.7198036424815655,
+      "adv": -9.503715409664437e-05,
+      "bbox": 0.9939679279923439,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.75,
+      "completions/max_terminated_length": 1592.75,
+      "completions/mean_length": 1573.19140625,
+      "completions/mean_terminated_length": 1573.19140625,
+      "completions/min_length": 1557.0,
+      "completions/min_terminated_length": 1557.0,
+      "entropy": 0.28550330083817244,
+      "epoch": 0.13267466478475654,
+      "geo": 0.9650512784719467,
+      "grad_norm": 0.359375,
+      "learning_rate": 8.165829145728644e-06,
+      "loss": 0.0004,
+      "num_tokens": 44444583.0,
+      "point_2d": 0.9892578125,
+      "point_3d": 0.83984375,
+      "reward": 0.8857182934880257,
+      "step": 376
+    },
+    {
+      "3d_bbox": 0.7569727748632431,
+      "adv": 0.0004161505894444417,
+      "bbox": 0.9042535312473774,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.5,
+      "completions/max_terminated_length": 1599.5,
+      "completions/mean_length": 1577.55859375,
+      "completions/mean_terminated_length": 1577.55859375,
+      "completions/min_length": 1561.375,
+      "completions/min_terminated_length": 1561.375,
+      "entropy": 0.2907078079879284,
+      "epoch": 0.13549752999294284,
+      "geo": 0.6764855124056339,
+      "grad_norm": 0.345703125,
+      "learning_rate": 8.125628140703518e-06,
+      "loss": 0.0001,
+      "num_tokens": 45389846.0,
+      "point_2d": 0.69140625,
+      "point_3d": 0.8193359375,
+      "reward": 0.7929921373724937,
+      "step": 384
+    },
+    {
+      "3d_bbox": 0.6764840334653854,
+      "adv": -1.4531655324390158e-05,
+      "bbox": 0.8646975085139275,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1600.5,
+      "completions/max_terminated_length": 1600.5,
+      "completions/mean_length": 1579.08984375,
+      "completions/mean_terminated_length": 1579.08984375,
+      "completions/min_length": 1558.375,
+      "completions/min_terminated_length": 1558.375,
+      "entropy": 0.2863780874758959,
+      "epoch": 0.13832039520112915,
+      "geo": 0.5404472178779542,
+      "grad_norm": 0.3125,
+      "learning_rate": 8.085427135678392e-06,
+      "loss": 0.0,
+      "num_tokens": 46335629.0,
+      "point_2d": 0.5048828125,
+      "point_3d": 0.806640625,
+      "reward": 0.713176254183054,
+      "step": 392
+    },
+    {
+      "3d_bbox": 0.7255214750766754,
+      "adv": -0.0006330321593850385,
+      "bbox": 0.891293590888381,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1602.375,
+      "completions/max_terminated_length": 1602.375,
+      "completions/mean_length": 1576.9921875,
+      "completions/mean_terminated_length": 1576.9921875,
+      "completions/min_length": 1558.125,
+      "completions/min_terminated_length": 1558.125,
+      "entropy": 0.28716066014021635,
+      "epoch": 0.14114326040931546,
+      "geo": 0.9741705134510994,
+      "grad_norm": 0.314453125,
+      "learning_rate": 8.045226130653267e-06,
+      "loss": 0.0004,
+      "num_tokens": 47280507.0,
+      "point_2d": 0.8916015625,
+      "point_3d": 0.859375,
+      "reward": 0.8419479243457317,
+      "step": 400
+    },
+    {
+      "3d_bbox": 0.7029357459396124,
+      "adv": 0.0008838025460136123,
+      "bbox": 0.9950641989707947,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.125,
+      "completions/max_terminated_length": 1598.125,
+      "completions/mean_length": 1575.92578125,
+      "completions/mean_terminated_length": 1575.92578125,
+      "completions/min_length": 1557.75,
+      "completions/min_terminated_length": 1557.75,
+      "entropy": 0.2837083409540355,
+      "epoch": 0.14396612561750177,
+      "geo": 0.864502564072609,
+      "grad_norm": 0.345703125,
+      "learning_rate": 8.005025125628141e-06,
+      "loss": -0.0003,
+      "num_tokens": 48225240.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8310546875,
+      "reward": 0.8822636753320694,
+      "step": 408
+    },
+    {
+      "3d_bbox": 0.6913789249956608,
+      "adv": -0.0002196824352722615,
+      "bbox": 0.9765095412731171,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.125,
+      "completions/max_terminated_length": 1598.125,
+      "completions/mean_length": 1579.28515625,
+      "completions/mean_terminated_length": 1579.28515625,
+      "completions/min_length": 1559.25,
+      "completions/min_terminated_length": 1559.25,
+      "entropy": 0.28565866220742464,
+      "epoch": 0.14678899082568808,
+      "geo": 0.8599895760416985,
+      "grad_norm": 0.30859375,
+      "learning_rate": 7.964824120603015e-06,
+      "loss": 0.0001,
+      "num_tokens": 49171057.0,
+      "point_2d": 1.0,
+      "point_3d": 0.904296875,
+      "reward": 0.8930463194847107,
+      "step": 416
+    },
+    {
+      "3d_bbox": 0.6788053698837757,
+      "adv": -0.00020617676273104735,
+      "bbox": 0.9748031497001648,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.25,
+      "completions/max_terminated_length": 1595.25,
+      "completions/mean_length": 1577.51953125,
+      "completions/mean_terminated_length": 1577.51953125,
+      "completions/min_length": 1559.625,
+      "completions/min_terminated_length": 1559.625,
+      "entropy": 0.28440852370113134,
+      "epoch": 0.1496118560338744,
+      "geo": 0.822394497692585,
+      "grad_norm": 0.40234375,
+      "learning_rate": 7.92462311557789e-06,
+      "loss": 0.0006,
+      "num_tokens": 50116302.0,
+      "point_2d": 0.875,
+      "point_3d": 0.8310546875,
+      "reward": 0.8399158045649529,
+      "step": 424
+    },
+    {
+      "3d_bbox": 0.6794650703668594,
+      "adv": 0.0011559298304746335,
+      "bbox": 0.9097459949553013,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.625,
+      "completions/max_terminated_length": 1596.625,
+      "completions/mean_length": 1573.80859375,
+      "completions/mean_terminated_length": 1573.80859375,
+      "completions/min_length": 1552.625,
+      "completions/min_terminated_length": 1552.625,
+      "entropy": 0.28376144589856267,
+      "epoch": 0.1524347212420607,
+      "geo": 0.7893084269016981,
+      "grad_norm": 0.3203125,
+      "learning_rate": 7.884422110552764e-06,
+      "loss": 0.0,
+      "num_tokens": 51060365.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8681640625,
+      "reward": 0.8643437847495079,
+      "step": 432
+    },
+    {
+      "3d_bbox": 0.7942164540290833,
+      "adv": -3.183347871527076e-05,
+      "bbox": 0.9926317036151886,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.75,
+      "completions/max_terminated_length": 1598.75,
+      "completions/mean_length": 1577.796875,
+      "completions/mean_terminated_length": 1577.796875,
+      "completions/min_length": 1553.125,
+      "completions/min_terminated_length": 1553.125,
+      "entropy": 0.2821194725111127,
+      "epoch": 0.155257586450247,
+      "geo": 0.9503362327814102,
+      "grad_norm": 0.3984375,
+      "learning_rate": 7.844221105527638e-06,
+      "loss": 0.0002,
+      "num_tokens": 52005633.0,
+      "point_2d": 1.0,
+      "point_3d": 0.869140625,
+      "reward": 0.9139972031116486,
+      "step": 440
+    },
+    {
+      "3d_bbox": 0.6752692721784115,
+      "adv": -9.695853077573702e-05,
+      "bbox": 0.982423759996891,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.25,
+      "completions/max_terminated_length": 1596.25,
+      "completions/mean_length": 1574.5546875,
+      "completions/mean_terminated_length": 1574.5546875,
+      "completions/min_length": 1555.125,
+      "completions/min_terminated_length": 1555.125,
+      "entropy": 0.2852726257406175,
+      "epoch": 0.15808045165843332,
+      "geo": 0.5512092066928744,
+      "grad_norm": 0.322265625,
+      "learning_rate": 7.804020100502512e-06,
+      "loss": -0.0004,
+      "num_tokens": 52949951.0,
+      "point_2d": 0.75,
+      "point_3d": 0.7998046875,
+      "reward": 0.801874428987503,
+      "step": 448
+    },
+    {
+      "3d_bbox": 0.7081510461866856,
+      "adv": 0.0003901398231391795,
+      "bbox": 0.9892760217189789,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.875,
+      "completions/max_terminated_length": 1596.875,
+      "completions/mean_length": 1573.0390625,
+      "completions/mean_terminated_length": 1573.0390625,
+      "completions/min_length": 1551.875,
+      "completions/min_terminated_length": 1551.875,
+      "entropy": 0.2837893716059625,
+      "epoch": 0.16090331686661963,
+      "geo": 0.9603376165032387,
+      "grad_norm": 0.3125,
+      "learning_rate": 7.763819095477387e-06,
+      "loss": -0.0007,
+      "num_tokens": 53894057.0,
+      "point_2d": 1.0,
+      "point_3d": 0.794921875,
+      "reward": 0.8730872422456741,
+      "step": 456
+    },
+    {
+      "3d_bbox": 0.7504811957478523,
+      "adv": 0.0007141208916436881,
+      "bbox": 0.9912561923265457,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.75,
+      "completions/max_terminated_length": 1594.75,
+      "completions/mean_length": 1576.1640625,
+      "completions/mean_terminated_length": 1576.1640625,
+      "completions/min_length": 1555.75,
+      "completions/min_terminated_length": 1555.75,
+      "entropy": 0.28930201614275575,
+      "epoch": 0.16372618207480594,
+      "geo": 0.8148366361856461,
+      "grad_norm": 0.275390625,
+      "learning_rate": 7.723618090452262e-06,
+      "loss": 0.0003,
+      "num_tokens": 54839099.0,
+      "point_2d": 0.875,
+      "point_3d": 0.8896484375,
+      "reward": 0.8765964508056641,
+      "step": 464
+    },
+    {
+      "3d_bbox": 0.819125022739172,
+      "adv": 0.00010633431156747974,
+      "bbox": 0.9906315729022026,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1603.0,
+      "completions/max_terminated_length": 1603.0,
+      "completions/mean_length": 1578.3203125,
+      "completions/mean_terminated_length": 1578.3203125,
+      "completions/min_length": 1556.375,
+      "completions/min_terminated_length": 1556.375,
+      "entropy": 0.28396829310804605,
+      "epoch": 0.16654904728299225,
+      "geo": 0.8597357729449868,
+      "grad_norm": 0.30078125,
+      "learning_rate": 7.683417085427137e-06,
+      "loss": 0.0004,
+      "num_tokens": 55784557.0,
+      "point_2d": 0.8916015625,
+      "point_3d": 0.9345703125,
+      "reward": 0.9089821055531502,
+      "step": 472
+    },
+    {
+      "3d_bbox": 0.7331585213541985,
+      "adv": -0.00033648866519797593,
+      "bbox": 0.8809062782675028,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.625,
+      "completions/max_terminated_length": 1594.625,
+      "completions/mean_length": 1576.11328125,
+      "completions/mean_terminated_length": 1576.11328125,
+      "completions/min_length": 1559.25,
+      "completions/min_terminated_length": 1559.25,
+      "entropy": 0.28669640561565757,
+      "epoch": 0.16937191249117856,
+      "geo": 0.9506581574678421,
+      "grad_norm": 0.2890625,
+      "learning_rate": 7.643216080402011e-06,
+      "loss": 0.0132,
+      "num_tokens": 56729514.0,
+      "point_2d": 0.8759765625,
+      "point_3d": 0.880859375,
+      "reward": 0.8427251875400543,
+      "step": 480
+    },
+    {
+      "3d_bbox": 0.7568283602595329,
+      "adv": -0.0008519564944435842,
+      "bbox": 0.9879242181777954,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1600.875,
+      "completions/max_terminated_length": 1600.875,
+      "completions/mean_length": 1574.87890625,
+      "completions/mean_terminated_length": 1574.87890625,
+      "completions/min_length": 1554.25,
+      "completions/min_terminated_length": 1554.25,
+      "entropy": 0.28737954096868634,
+      "epoch": 0.17219477769936486,
+      "geo": 0.7571075335144997,
+      "grad_norm": 0.400390625,
+      "learning_rate": 7.603015075376885e-06,
+      "loss": 0.0105,
+      "num_tokens": 57674099.0,
+      "point_2d": 0.7998046875,
+      "point_3d": 0.9228515625,
+      "reward": 0.8668522015213966,
+      "step": 488
+    },
+    {
+      "3d_bbox": 0.7658256888389587,
+      "adv": 0.0011402575073589105,
+      "bbox": 0.9932135492563248,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.25,
+      "completions/max_terminated_length": 1596.25,
+      "completions/mean_length": 1574.265625,
+      "completions/mean_terminated_length": 1574.265625,
+      "completions/min_length": 1554.375,
+      "completions/min_terminated_length": 1554.375,
+      "entropy": 0.2865150640718639,
+      "epoch": 0.17501764290755117,
+      "geo": 0.8641151636838913,
+      "grad_norm": 0.37890625,
+      "learning_rate": 7.5628140703517595e-06,
+      "loss": -0.0004,
+      "num_tokens": 58618575.0,
+      "point_2d": 0.91015625,
+      "point_3d": 0.904296875,
+      "reward": 0.8933730870485306,
+      "step": 496
+    },
+    {
+      "3d_bbox": 0.7145381905138493,
+      "adv": 0.0003104716488451231,
+      "bbox": 0.9926531538367271,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.75,
+      "completions/max_terminated_length": 1594.75,
+      "completions/mean_length": 1573.12890625,
+      "completions/mean_terminated_length": 1573.12890625,
+      "completions/min_length": 1553.375,
+      "completions/min_terminated_length": 1553.375,
+      "entropy": 0.28603550279513,
+      "epoch": 0.17784050811573748,
+      "geo": 0.7219396531581879,
+      "grad_norm": 0.310546875,
+      "learning_rate": 7.522613065326634e-06,
+      "loss": -0.0,
+      "num_tokens": 59562552.0,
+      "point_2d": 0.7509765625,
+      "point_3d": 0.9365234375,
+      "reward": 0.8486728444695473,
+      "step": 504
+    },
+    {
+      "3d_bbox": 0.7242622300982475,
+      "adv": -0.0005710294717573561,
+      "bbox": 0.9246156923472881,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1591.75,
+      "completions/max_terminated_length": 1591.75,
+      "completions/mean_length": 1571.015625,
+      "completions/mean_terminated_length": 1571.015625,
+      "completions/min_length": 1551.625,
+      "completions/min_terminated_length": 1551.625,
+      "entropy": 0.2915982282720506,
+      "epoch": 0.1806633733239238,
+      "geo": 0.7993985265493393,
+      "grad_norm": 0.408203125,
+      "learning_rate": 7.482412060301508e-06,
+      "loss": 0.0001,
+      "num_tokens": 60505948.0,
+      "point_2d": 0.7919921875,
+      "point_3d": 0.8671875,
+      "reward": 0.8270144015550613,
+      "step": 512
+    },
+    {
+      "3d_bbox": 0.7831042557954788,
+      "adv": -0.000159245344093506,
+      "bbox": 0.9960433393716812,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.375,
+      "completions/max_terminated_length": 1592.375,
+      "completions/mean_length": 1574.84375,
+      "completions/mean_terminated_length": 1574.84375,
+      "completions/min_length": 1559.125,
+      "completions/min_terminated_length": 1559.125,
+      "entropy": 0.2845937805250287,
+      "epoch": 0.1834862385321101,
+      "geo": 0.9170719608664513,
+      "grad_norm": 0.345703125,
+      "learning_rate": 7.442211055276382e-06,
+      "loss": -0.0,
+      "num_tokens": 61450564.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8408203125,
+      "reward": 0.90499197691679,
+      "step": 520
+    },
+    {
+      "3d_bbox": 0.7821980714797974,
+      "adv": -0.000531579293010509,
+      "bbox": 0.9920122027397156,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.375,
+      "completions/max_terminated_length": 1595.375,
+      "completions/mean_length": 1571.765625,
+      "completions/mean_terminated_length": 1571.765625,
+      "completions/min_length": 1551.5,
+      "completions/min_terminated_length": 1551.5,
+      "entropy": 0.2882155287079513,
+      "epoch": 0.1863091037402964,
+      "geo": 0.901853509247303,
+      "grad_norm": 0.365234375,
+      "learning_rate": 7.402010050251257e-06,
+      "loss": 0.0002,
+      "num_tokens": 62394336.0,
+      "point_2d": 0.888671875,
+      "point_3d": 0.9052734375,
+      "reward": 0.8920388966798782,
+      "step": 528
+    },
+    {
+      "3d_bbox": 0.7537402808666229,
+      "adv": -0.0004438982141437009,
+      "bbox": 0.9946581870317459,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.375,
+      "completions/max_terminated_length": 1594.375,
+      "completions/mean_length": 1574.3203125,
+      "completions/mean_terminated_length": 1574.3203125,
+      "completions/min_length": 1554.75,
+      "completions/min_terminated_length": 1554.75,
+      "entropy": 0.2886669235303998,
+      "epoch": 0.18913196894848272,
+      "geo": 0.8310712724924088,
+      "grad_norm": 0.36328125,
+      "learning_rate": 7.361809045226132e-06,
+      "loss": 0.0002,
+      "num_tokens": 63338834.0,
+      "point_2d": 0.873046875,
+      "point_3d": 0.904296875,
+      "reward": 0.8814355507493019,
+      "step": 536
+    },
+    {
+      "3d_bbox": 0.6960996612906456,
+      "adv": 0.0006180387658787367,
+      "bbox": 0.869446316966787,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1597.25,
+      "completions/max_terminated_length": 1597.25,
+      "completions/mean_length": 1571.88671875,
+      "completions/mean_terminated_length": 1571.88671875,
+      "completions/min_length": 1552.25,
+      "completions/min_terminated_length": 1552.25,
+      "entropy": 0.2887332271784544,
+      "epoch": 0.19195483415666903,
+      "geo": 0.826970137655735,
+      "grad_norm": 0.267578125,
+      "learning_rate": 7.321608040201006e-06,
+      "loss": -0.0,
+      "num_tokens": 64282685.0,
+      "point_2d": 0.875,
+      "point_3d": 0.8662109375,
+      "reward": 0.8266892246901989,
+      "step": 544
+    },
+    {
+      "3d_bbox": 0.8654715269804001,
+      "adv": -0.0008252508814621251,
+      "bbox": 0.9932455867528915,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.125,
+      "completions/max_terminated_length": 1596.125,
+      "completions/mean_length": 1572.1953125,
+      "completions/mean_terminated_length": 1572.1953125,
+      "completions/min_length": 1553.625,
+      "completions/min_terminated_length": 1553.625,
+      "entropy": 0.2873402568511665,
+      "epoch": 0.19477769936485534,
+      "geo": 0.872327484190464,
+      "grad_norm": 0.2890625,
+      "learning_rate": 7.28140703517588e-06,
+      "loss": 0.0004,
+      "num_tokens": 65226639.0,
+      "point_2d": 0.8935546875,
+      "point_3d": 0.9482421875,
+      "reward": 0.9251285046339035,
+      "step": 552
+    },
+    {
+      "3d_bbox": 0.7851101011037827,
+      "adv": -0.0011301148442726117,
+      "bbox": 0.9379025921225548,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1588.5,
+      "completions/max_terminated_length": 1588.5,
+      "completions/mean_length": 1571.4765625,
+      "completions/mean_terminated_length": 1571.4765625,
+      "completions/min_length": 1555.0,
+      "completions/min_terminated_length": 1555.0,
+      "entropy": 0.2852159277535975,
+      "epoch": 0.19760056457304165,
+      "geo": 0.690475556999445,
+      "grad_norm": 0.3125,
+      "learning_rate": 7.241206030150754e-06,
+      "loss": 0.0006,
+      "num_tokens": 66170305.0,
+      "point_2d": 0.625,
+      "point_3d": 0.837890625,
+      "reward": 0.7964758351445198,
+      "step": 560
+    },
+    {
+      "3d_bbox": 0.7313521131873131,
+      "adv": -0.000422738248744281,
+      "bbox": 0.8720479012117721,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.875,
+      "completions/max_terminated_length": 1601.875,
+      "completions/mean_length": 1574.91015625,
+      "completions/mean_terminated_length": 1574.91015625,
+      "completions/min_length": 1551.25,
+      "completions/min_terminated_length": 1551.25,
+      "entropy": 0.2880417066626251,
+      "epoch": 0.20042342978122796,
+      "geo": 0.8383861493784934,
+      "grad_norm": 0.412109375,
+      "learning_rate": 7.2010050251256295e-06,
+      "loss": 0.0004,
+      "num_tokens": 67114842.0,
+      "point_2d": 0.8740234375,
+      "point_3d": 0.84765625,
+      "reward": 0.8312699124217033,
+      "step": 568
+    },
+    {
+      "3d_bbox": 0.6841592267155647,
+      "adv": -0.0008141829575833981,
+      "bbox": 0.9068506434559822,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1597.5,
+      "completions/max_terminated_length": 1597.5,
+      "completions/mean_length": 1571.79296875,
+      "completions/mean_terminated_length": 1571.79296875,
+      "completions/min_length": 1546.75,
+      "completions/min_terminated_length": 1546.75,
+      "entropy": 0.2889366364106536,
+      "epoch": 0.20324629498941427,
+      "geo": 0.7308010272681713,
+      "grad_norm": 0.2578125,
+      "learning_rate": 7.160804020100504e-06,
+      "loss": 0.0002,
+      "num_tokens": 68058501.0,
+      "point_2d": 0.8525390625,
+      "point_3d": 0.794921875,
+      "reward": 0.8096176832914352,
+      "step": 576
+    },
+    {
+      "3d_bbox": 0.5905975513160229,
+      "adv": -0.00022596626513404772,
+      "bbox": 0.9184868931770325,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1597.5,
+      "completions/max_terminated_length": 1597.5,
+      "completions/mean_length": 1572.45703125,
+      "completions/mean_terminated_length": 1572.45703125,
+      "completions/min_length": 1551.125,
+      "completions/min_terminated_length": 1551.125,
+      "entropy": 0.2837987565435469,
+      "epoch": 0.20606916019760058,
+      "geo": 0.8267702832818031,
+      "grad_norm": 0.314453125,
+      "learning_rate": 7.120603015075378e-06,
+      "loss": -0.0003,
+      "num_tokens": 69002250.0,
+      "point_2d": 0.8681640625,
+      "point_3d": 0.7445197626948357,
+      "reward": 0.7804420664906502,
+      "step": 584
+    },
+    {
+      "3d_bbox": 0.6897582598030567,
+      "adv": -0.0004831653022847604,
+      "bbox": 0.9742325842380524,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.25,
+      "completions/max_terminated_length": 1595.25,
+      "completions/mean_length": 1572.9296875,
+      "completions/mean_terminated_length": 1572.9296875,
+      "completions/min_length": 1553.5,
+      "completions/min_terminated_length": 1553.5,
+      "entropy": 0.2835669592022896,
+      "epoch": 0.20889202540578689,
+      "geo": 0.8361655548214912,
+      "grad_norm": 0.330078125,
+      "learning_rate": 7.080402010050251e-06,
+      "loss": 0.0004,
+      "num_tokens": 69946400.0,
+      "point_2d": 0.875,
+      "point_3d": 0.8948184698820114,
+      "reward": 0.8584523350000381,
+      "step": 592
+    },
+    {
+      "3d_bbox": 0.785271767526865,
+      "adv": -0.00028875649149995297,
+      "bbox": 0.9752914533019066,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.0,
+      "completions/max_terminated_length": 1599.0,
+      "completions/mean_length": 1572.3671875,
+      "completions/mean_terminated_length": 1572.3671875,
+      "completions/min_length": 1549.875,
+      "completions/min_terminated_length": 1549.875,
+      "entropy": 0.29118923749774694,
+      "epoch": 0.2117148906139732,
+      "geo": 0.9338327869772911,
+      "grad_norm": 0.3671875,
+      "learning_rate": 7.040201005025126e-06,
+      "loss": 0.0007,
+      "num_tokens": 70890342.0,
+      "point_2d": 0.998046875,
+      "point_3d": 0.94140625,
+      "reward": 0.9250040724873543,
+      "step": 600
+    },
+    {
+      "3d_bbox": 0.8766504302620888,
+      "adv": -0.002171848598663928,
+      "bbox": 0.9872715324163437,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.125,
+      "completions/max_terminated_length": 1601.125,
+      "completions/mean_length": 1576.87890625,
+      "completions/mean_terminated_length": 1576.87890625,
+      "completions/min_length": 1556.375,
+      "completions/min_terminated_length": 1556.375,
+      "entropy": 0.28933983854949474,
+      "epoch": 0.2145377558221595,
+      "geo": 0.845995269715786,
+      "grad_norm": 0.33203125,
+      "learning_rate": 7e-06,
+      "loss": 0.0003,
+      "num_tokens": 71835359.0,
+      "point_2d": 0.875,
+      "point_3d": 0.9454848319292068,
+      "reward": 0.9211016893386841,
+      "step": 608
+    },
+    {
+      "3d_bbox": 0.8059130385518074,
+      "adv": 0.0005476666337926872,
+      "bbox": 0.9907352328300476,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1593.625,
+      "completions/max_terminated_length": 1593.625,
+      "completions/mean_length": 1570.9921875,
+      "completions/mean_terminated_length": 1570.9921875,
+      "completions/min_length": 1554.125,
+      "completions/min_terminated_length": 1554.125,
+      "entropy": 0.28401868836954236,
+      "epoch": 0.2173606210303458,
+      "geo": 0.8720198646187782,
+      "grad_norm": 0.384765625,
+      "learning_rate": 6.959798994974874e-06,
+      "loss": -0.0004,
+      "num_tokens": 72778933.0,
+      "point_2d": 0.9248046875,
+      "point_3d": 0.9135454967617989,
+      "reward": 0.9087496176362038,
+      "step": 616
+    },
+    {
+      "3d_bbox": 0.8549554497003555,
+      "adv": -0.0005420902471087174,
+      "bbox": 0.9758080616593361,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1593.0,
+      "completions/max_terminated_length": 1593.0,
+      "completions/mean_length": 1570.1640625,
+      "completions/mean_terminated_length": 1570.1640625,
+      "completions/min_length": 1549.0,
+      "completions/min_terminated_length": 1549.0,
+      "entropy": 0.29045473085716367,
+      "epoch": 0.22018348623853212,
+      "geo": 0.8916067183017731,
+      "grad_norm": 0.3828125,
+      "learning_rate": 6.919597989949749e-06,
+      "loss": 0.0156,
+      "num_tokens": 73722191.0,
+      "point_2d": 1.0,
+      "point_3d": 0.9639246314764023,
+      "reward": 0.9486720338463783,
+      "step": 624
+    },
+    {
+      "3d_bbox": 0.7975610345602036,
+      "adv": 0.0014872762099003012,
+      "bbox": 0.9193930514156818,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1600.625,
+      "completions/max_terminated_length": 1600.625,
+      "completions/mean_length": 1573.29296875,
+      "completions/mean_terminated_length": 1573.29296875,
+      "completions/min_length": 1553.0,
+      "completions/min_terminated_length": 1553.0,
+      "entropy": 0.28542913123965263,
+      "epoch": 0.22300635144671843,
+      "geo": 0.936805784702301,
+      "grad_norm": 0.2734375,
+      "learning_rate": 6.8793969849246235e-06,
+      "loss": 0.0152,
+      "num_tokens": 74666354.0,
+      "point_2d": 0.9990234375,
+      "point_3d": 0.93359375,
+      "reward": 0.9123928248882294,
+      "step": 632
+    },
+    {
+      "3d_bbox": 0.6216618679463863,
+      "adv": -0.0013801682944176719,
+      "bbox": 0.9884917959570885,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.375,
+      "completions/max_terminated_length": 1601.375,
+      "completions/mean_length": 1577.00390625,
+      "completions/mean_terminated_length": 1577.00390625,
+      "completions/min_length": 1554.0,
+      "completions/min_terminated_length": 1554.0,
+      "entropy": 0.28678226098418236,
+      "epoch": 0.22582921665490474,
+      "geo": 0.9096385985612869,
+      "grad_norm": 0.42578125,
+      "learning_rate": 6.839195979899498e-06,
+      "loss": 0.0005,
+      "num_tokens": 75611339.0,
+      "point_2d": 0.8291015625,
+      "point_3d": 0.7913028448820114,
+      "reward": 0.8076395243406296,
+      "step": 640
+    },
+    {
+      "3d_bbox": 0.7582510337233543,
+      "adv": -0.00019565537664334443,
+      "bbox": 0.9911796301603317,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.625,
+      "completions/max_terminated_length": 1596.625,
+      "completions/mean_length": 1575.71484375,
+      "completions/mean_terminated_length": 1575.71484375,
+      "completions/min_length": 1556.25,
+      "completions/min_terminated_length": 1556.25,
+      "entropy": 0.2863654876127839,
+      "epoch": 0.22865208186309105,
+      "geo": 0.9025658443570137,
+      "grad_norm": 0.451171875,
+      "learning_rate": 6.798994974874372e-06,
+      "loss": -0.0001,
+      "num_tokens": 76556042.0,
+      "point_2d": 0.95703125,
+      "point_3d": 0.8994140625,
+      "reward": 0.9014689922332764,
+      "step": 648
+    },
+    {
+      "3d_bbox": 0.753610834479332,
+      "adv": 0.00039293405916396296,
+      "bbox": 0.9021090529859066,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.625,
+      "completions/max_terminated_length": 1592.625,
+      "completions/mean_length": 1572.203125,
+      "completions/mean_terminated_length": 1572.203125,
+      "completions/min_length": 1552.25,
+      "completions/min_terminated_length": 1552.25,
+      "entropy": 0.286099414806813,
+      "epoch": 0.23147494707127736,
+      "geo": 0.9557241797447205,
+      "grad_norm": 0.455078125,
+      "learning_rate": 6.758793969849246e-06,
+      "loss": 0.0002,
+      "num_tokens": 77499886.0,
+      "point_2d": 1.0,
+      "point_3d": 0.91015625,
+      "reward": 0.8914690390229225,
+      "step": 656
+    },
+    {
+      "3d_bbox": 0.831758588552475,
+      "adv": -7.158980952226557e-05,
+      "bbox": 0.9794979766011238,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.5,
+      "completions/max_terminated_length": 1595.5,
+      "completions/mean_length": 1574.5,
+      "completions/mean_terminated_length": 1574.5,
+      "completions/min_length": 1558.75,
+      "completions/min_terminated_length": 1558.75,
+      "entropy": 0.2819104827940464,
+      "epoch": 0.23429781227946367,
+      "geo": 0.8940749987959862,
+      "grad_norm": 0.294921875,
+      "learning_rate": 6.718592964824121e-06,
+      "loss": 0.0001,
+      "num_tokens": 78444526.0,
+      "point_2d": 1.0,
+      "point_3d": 0.939453125,
+      "reward": 0.9376774355769157,
+      "step": 664
+    },
+    {
+      "3d_bbox": 0.8210256397724152,
+      "adv": 0.000540634344361024,
+      "bbox": 0.9576755836606026,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.125,
+      "completions/max_terminated_length": 1594.125,
+      "completions/mean_length": 1573.80859375,
+      "completions/mean_terminated_length": 1573.80859375,
+      "completions/min_length": 1553.25,
+      "completions/min_terminated_length": 1553.25,
+      "entropy": 0.2863646815530956,
+      "epoch": 0.23712067748764998,
+      "geo": 0.9599328562617302,
+      "grad_norm": 0.28125,
+      "learning_rate": 6.678391959798996e-06,
+      "loss": -0.0001,
+      "num_tokens": 79388949.0,
+      "point_2d": 1.0,
+      "point_3d": 0.92578125,
+      "reward": 0.9261206015944481,
+      "step": 672
+    },
+    {
+      "3d_bbox": 0.8162731677293777,
+      "adv": -0.0006006620242260396,
+      "bbox": 0.9639560282230377,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1593.75,
+      "completions/max_terminated_length": 1593.75,
+      "completions/mean_length": 1573.62890625,
+      "completions/mean_terminated_length": 1573.62890625,
+      "completions/min_length": 1557.125,
+      "completions/min_terminated_length": 1557.125,
+      "entropy": 0.2868205299600959,
+      "epoch": 0.2399435426958363,
+      "geo": 0.9127583503723145,
+      "grad_norm": 0.44921875,
+      "learning_rate": 6.63819095477387e-06,
+      "loss": -0.0,
+      "num_tokens": 80333190.0,
+      "point_2d": 1.0,
+      "point_3d": 0.9267578125,
+      "reward": 0.926746778190136,
+      "step": 680
+    },
+    {
+      "3d_bbox": 0.7580128684639931,
+      "adv": 0.0009847661577850886,
+      "bbox": 0.9946200400590897,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.0,
+      "completions/max_terminated_length": 1598.0,
+      "completions/mean_length": 1572.2109375,
+      "completions/mean_terminated_length": 1572.2109375,
+      "completions/min_length": 1552.0,
+      "completions/min_terminated_length": 1552.0,
+      "entropy": 0.286160247400403,
+      "epoch": 0.2427664079040226,
+      "geo": 0.9191197454929352,
+      "grad_norm": 0.361328125,
+      "learning_rate": 6.597989949748744e-06,
+      "loss": 0.0001,
+      "num_tokens": 81277036.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8818359375,
+      "reward": 0.9086171984672546,
+      "step": 688
+    },
+    {
+      "3d_bbox": 0.729152224957943,
+      "adv": -0.0009101523137360346,
+      "bbox": 0.9927223324775696,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.875,
+      "completions/max_terminated_length": 1592.875,
+      "completions/mean_length": 1571.8515625,
+      "completions/mean_terminated_length": 1571.8515625,
+      "completions/min_length": 1553.375,
+      "completions/min_terminated_length": 1553.375,
+      "entropy": 0.28340971609577537,
+      "epoch": 0.2455892731122089,
+      "geo": 0.9836970791220665,
+      "grad_norm": 0.34765625,
+      "learning_rate": 6.557788944723618e-06,
+      "loss": 0.0005,
+      "num_tokens": 82220822.0,
+      "point_2d": 1.0,
+      "point_3d": 0.9014246314764023,
+      "reward": 0.9058247655630112,
+      "step": 696
+    },
+    {
+      "3d_bbox": 0.7259362824261189,
+      "adv": 0.0003162266129947966,
+      "bbox": 0.922937985509634,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.0,
+      "completions/max_terminated_length": 1596.0,
+      "completions/mean_length": 1575.2265625,
+      "completions/mean_terminated_length": 1575.2265625,
+      "completions/min_length": 1555.625,
+      "completions/min_terminated_length": 1555.625,
+      "entropy": 0.286905855871737,
+      "epoch": 0.24841213832039521,
+      "geo": 0.8376269713044167,
+      "grad_norm": 0.30078125,
+      "learning_rate": 6.5175879396984935e-06,
+      "loss": 0.0,
+      "num_tokens": 83165600.0,
+      "point_2d": 0.9541015625,
+      "point_3d": 0.8759765625,
+      "reward": 0.8697381019592285,
+      "step": 704
+    },
+    {
+      "3d_bbox": 0.7864346951246262,
+      "adv": -0.0007161469466154813,
+      "bbox": 0.9958559349179268,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.0,
+      "completions/max_terminated_length": 1595.0,
+      "completions/mean_length": 1572.4921875,
+      "completions/mean_terminated_length": 1572.4921875,
+      "completions/min_length": 1545.125,
+      "completions/min_terminated_length": 1545.125,
+      "entropy": 0.2886675829067826,
+      "epoch": 0.2512350035285815,
+      "geo": 0.8416303154081106,
+      "grad_norm": 0.333984375,
+      "learning_rate": 6.477386934673368e-06,
+      "loss": 0.0008,
+      "num_tokens": 84109606.0,
+      "point_2d": 0.75,
+      "point_3d": 0.947265625,
+      "reward": 0.8698890581727028,
+      "step": 712
+    },
+    {
+      "3d_bbox": 0.657585222274065,
+      "adv": -0.0004964938416378573,
+      "bbox": 0.9940767288208008,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1589.75,
+      "completions/max_terminated_length": 1589.75,
+      "completions/mean_length": 1571.515625,
+      "completions/mean_terminated_length": 1571.515625,
+      "completions/min_length": 1553.5,
+      "completions/min_terminated_length": 1553.5,
+      "entropy": 0.2854380081407726,
+      "epoch": 0.25405786873676783,
+      "geo": 0.8985359221696854,
+      "grad_norm": 0.275390625,
+      "learning_rate": 6.437185929648242e-06,
+      "loss": 0.0,
+      "num_tokens": 85053346.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8623046875,
+      "reward": 0.8784916624426842,
+      "step": 720
+    },
+    {
+      "3d_bbox": 0.8299823924899101,
+      "adv": -0.0006786447083868552,
+      "bbox": 0.9957724660634995,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.75,
+      "completions/max_terminated_length": 1599.75,
+      "completions/mean_length": 1575.05078125,
+      "completions/mean_terminated_length": 1575.05078125,
+      "completions/min_length": 1553.625,
+      "completions/min_terminated_length": 1553.625,
+      "entropy": 0.2834048862569034,
+      "epoch": 0.25688073394495414,
+      "geo": 0.7720311819575727,
+      "grad_norm": 0.353515625,
+      "learning_rate": 6.396984924623116e-06,
+      "loss": 0.0008,
+      "num_tokens": 85997743.0,
+      "point_2d": 0.875,
+      "point_3d": 0.8583984375,
+      "reward": 0.8897883221507072,
+      "step": 728
+    },
+    {
+      "3d_bbox": 0.814798042178154,
+      "adv": -0.00043147145015609567,
+      "bbox": 0.994331993162632,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1603.0,
+      "completions/max_terminated_length": 1603.0,
+      "completions/mean_length": 1578.44921875,
+      "completions/mean_terminated_length": 1578.44921875,
+      "completions/min_length": 1556.25,
+      "completions/min_terminated_length": 1556.25,
+      "entropy": 0.2801376706920564,
+      "epoch": 0.25970359915314045,
+      "geo": 0.9146922156214714,
+      "grad_norm": 0.25,
+      "learning_rate": 6.3567839195979905e-06,
+      "loss": 0.0001,
+      "num_tokens": 86943082.0,
+      "point_2d": 0.8798828125,
+      "point_3d": 0.94921875,
+      "reward": 0.909557893872261,
+      "step": 736
+    },
+    {
+      "3d_bbox": 0.7457513771951199,
+      "adv": -1.4886696476423822e-05,
+      "bbox": 0.9952592924237251,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.875,
+      "completions/max_terminated_length": 1592.875,
+      "completions/mean_length": 1572.21484375,
+      "completions/mean_terminated_length": 1572.21484375,
+      "completions/min_length": 1553.625,
+      "completions/min_terminated_length": 1553.625,
+      "entropy": 0.28585455380380154,
+      "epoch": 0.26252646436132676,
+      "geo": 0.9565036073327065,
+      "grad_norm": 0.345703125,
+      "learning_rate": 6.316582914572866e-06,
+      "loss": -0.0003,
+      "num_tokens": 87886913.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8095703125,
+      "reward": 0.8876452520489693,
+      "step": 744
+    },
+    {
+      "3d_bbox": 0.6948670484125614,
+      "adv": -0.00016846299786266172,
+      "bbox": 0.8553073853254318,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.375,
+      "completions/max_terminated_length": 1594.375,
+      "completions/mean_length": 1571.93359375,
+      "completions/mean_terminated_length": 1571.93359375,
+      "completions/min_length": 1552.0,
+      "completions/min_terminated_length": 1552.0,
+      "entropy": 0.28335432009771466,
+      "epoch": 0.26534932956951307,
+      "geo": 0.7740265503525734,
+      "grad_norm": 0.421875,
+      "learning_rate": 6.27638190954774e-06,
+      "loss": 0.0162,
+      "num_tokens": 88830776.0,
+      "point_2d": 0.8701171875,
+      "point_3d": 0.9082605689764023,
+      "reward": 0.8321380317211151,
+      "step": 752
+    },
+    {
+      "3d_bbox": 0.7499241754412651,
+      "adv": 0.0011103415663455962,
+      "bbox": 0.9951865822076797,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.75,
+      "completions/max_terminated_length": 1592.75,
+      "completions/mean_length": 1570.93359375,
+      "completions/mean_terminated_length": 1570.93359375,
+      "completions/min_length": 1552.25,
+      "completions/min_terminated_length": 1552.25,
+      "entropy": 0.28731182869523764,
+      "epoch": 0.2681721947776994,
+      "geo": 0.620879240334034,
+      "grad_norm": 0.2197265625,
+      "learning_rate": 6.236180904522614e-06,
+      "loss": 0.0001,
+      "num_tokens": 89774255.0,
+      "point_2d": 0.7119140625,
+      "point_3d": 0.8916015625,
+      "reward": 0.8371566012501717,
+      "step": 760
+    },
+    {
+      "3d_bbox": 0.7572767026722431,
+      "adv": 0.0006080263119656593,
+      "bbox": 0.9954995736479759,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1589.0,
+      "completions/max_terminated_length": 1589.0,
+      "completions/mean_length": 1569.015625,
+      "completions/mean_terminated_length": 1569.015625,
+      "completions/min_length": 1549.75,
+      "completions/min_terminated_length": 1549.75,
+      "entropy": 0.2817346737720072,
+      "epoch": 0.2709950599858857,
+      "geo": 0.8461784943938255,
+      "grad_norm": 0.41015625,
+      "learning_rate": 6.195979899497488e-06,
+      "loss": -0.0,
+      "num_tokens": 90717243.0,
+      "point_2d": 1.0,
+      "point_3d": 0.892578125,
+      "reward": 0.9113386049866676,
+      "step": 768
+    },
+    {
+      "3d_bbox": 0.8124237954616547,
+      "adv": -0.00010392009426141158,
+      "bbox": 0.9695260003209114,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1590.0,
+      "completions/max_terminated_length": 1590.0,
+      "completions/mean_length": 1571.890625,
+      "completions/mean_terminated_length": 1571.890625,
+      "completions/min_length": 1553.75,
+      "completions/min_terminated_length": 1553.75,
+      "entropy": 0.2867649751715362,
+      "epoch": 0.273817925194072,
+      "geo": 0.7172370981425047,
+      "grad_norm": 0.29296875,
+      "learning_rate": 6.155778894472362e-06,
+      "loss": -0.0,
+      "num_tokens": 91660903.0,
+      "point_2d": 0.75,
+      "point_3d": 0.9462890625,
+      "reward": 0.8695597127079964,
+      "step": 776
+    },
+    {
+      "3d_bbox": 0.6247975081205368,
+      "adv": 0.00014501793702947907,
+      "bbox": 0.9793029651045799,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.125,
+      "completions/max_terminated_length": 1592.125,
+      "completions/mean_length": 1571.03515625,
+      "completions/mean_terminated_length": 1571.03515625,
+      "completions/min_length": 1550.875,
+      "completions/min_terminated_length": 1550.875,
+      "entropy": 0.2792881946079433,
+      "epoch": 0.2766407904022583,
+      "geo": 0.8185150865465403,
+      "grad_norm": 0.2890625,
+      "learning_rate": 6.115577889447236e-06,
+      "loss": 0.0,
+      "num_tokens": 92604352.0,
+      "point_2d": 0.8876953125,
+      "point_3d": 0.8994140625,
+      "reward": 0.8478024452924728,
+      "step": 784
+    },
+    {
+      "3d_bbox": 0.7522557079792023,
+      "adv": 0.00011037173771910602,
+      "bbox": 0.8692253082990646,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.625,
+      "completions/max_terminated_length": 1594.625,
+      "completions/mean_length": 1572.27734375,
+      "completions/mean_terminated_length": 1572.27734375,
+      "completions/min_length": 1552.5,
+      "completions/min_terminated_length": 1552.5,
+      "entropy": 0.2874673060141504,
+      "epoch": 0.2794636556104446,
+      "geo": 0.9408122524619102,
+      "grad_norm": 0.31640625,
+      "learning_rate": 6.07537688442211e-06,
+      "loss": 0.0144,
+      "num_tokens": 93548159.0,
+      "point_2d": 0.7587890625,
+      "point_3d": 0.91796875,
+      "reward": 0.8245597034692764,
+      "step": 792
+    },
+    {
+      "3d_bbox": 0.8205922544002533,
+      "adv": 0.0001977139909286052,
+      "bbox": 0.9971112534403801,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1589.25,
+      "completions/max_terminated_length": 1589.25,
+      "completions/mean_length": 1571.43359375,
+      "completions/mean_terminated_length": 1571.43359375,
+      "completions/min_length": 1552.625,
+      "completions/min_terminated_length": 1552.625,
+      "entropy": 0.28211602522060275,
+      "epoch": 0.2822865208186309,
+      "geo": 0.8445243984460831,
+      "grad_norm": 0.3203125,
+      "learning_rate": 6.035175879396985e-06,
+      "loss": 0.0007,
+      "num_tokens": 94491662.0,
+      "point_2d": 0.884765625,
+      "point_3d": 0.9365234375,
+      "reward": 0.9097481444478035,
+      "step": 800
+    },
+    {
+      "3d_bbox": 0.794718824326992,
+      "adv": -0.00023952151968842372,
+      "bbox": 0.9944768622517586,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.375,
+      "completions/max_terminated_length": 1601.375,
+      "completions/mean_length": 1571.67578125,
+      "completions/mean_terminated_length": 1571.67578125,
+      "completions/min_length": 1552.125,
+      "completions/min_terminated_length": 1552.125,
+      "entropy": 0.2849376816302538,
+      "epoch": 0.28510938602681724,
+      "geo": 0.9431384205818176,
+      "grad_norm": 0.34375,
+      "learning_rate": 5.99497487437186e-06,
+      "loss": 0.0001,
+      "num_tokens": 95435451.0,
+      "point_2d": 0.9990234375,
+      "point_3d": 0.8720703125,
+      "reward": 0.9150723442435265,
+      "step": 808
+    },
+    {
+      "3d_bbox": 0.7696447446942329,
+      "adv": 0.0008425721316598356,
+      "bbox": 0.9485463574528694,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1593.875,
+      "completions/max_terminated_length": 1593.875,
+      "completions/mean_length": 1570.11328125,
+      "completions/mean_terminated_length": 1570.11328125,
+      "completions/min_length": 1553.875,
+      "completions/min_terminated_length": 1553.875,
+      "entropy": 0.28682070411741734,
+      "epoch": 0.28793225123500354,
+      "geo": 0.9097637981176376,
+      "grad_norm": 0.2353515625,
+      "learning_rate": 5.954773869346734e-06,
+      "loss": -0.0006,
+      "num_tokens": 96378680.0,
+      "point_2d": 0.9990234375,
+      "point_3d": 0.9013671875,
+      "reward": 0.9046454280614853,
+      "step": 816
+    },
+    {
+      "3d_bbox": 0.6897615268826485,
+      "adv": -0.000497654465107189,
+      "bbox": 0.8617971315979958,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1602.125,
+      "completions/max_terminated_length": 1602.125,
+      "completions/mean_length": 1577.7734375,
+      "completions/mean_terminated_length": 1577.7734375,
+      "completions/min_length": 1558.125,
+      "completions/min_terminated_length": 1558.125,
+      "entropy": 0.28228796226903796,
+      "epoch": 0.29075511644318985,
+      "geo": 0.7427511867135763,
+      "grad_norm": 0.326171875,
+      "learning_rate": 5.914572864321608e-06,
+      "loss": 0.0001,
+      "num_tokens": 97324310.0,
+      "point_2d": 0.7529296875,
+      "point_3d": 0.919921875,
+      "reward": 0.8061025515198708,
+      "step": 824
+    },
+    {
+      "3d_bbox": 0.6537492573261261,
+      "adv": -0.00023956477116371389,
+      "bbox": 0.9781710132956505,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.0,
+      "completions/max_terminated_length": 1596.0,
+      "completions/mean_length": 1574.296875,
+      "completions/mean_terminated_length": 1574.296875,
+      "completions/min_length": 1554.75,
+      "completions/min_terminated_length": 1554.75,
+      "entropy": 0.28043507505208254,
+      "epoch": 0.29357798165137616,
+      "geo": 0.8402111828327179,
+      "grad_norm": 0.291015625,
+      "learning_rate": 5.874371859296482e-06,
+      "loss": -0.0001,
+      "num_tokens": 98268682.0,
+      "point_2d": 1.0,
+      "point_3d": 0.9551355689764023,
+      "reward": 0.8967639729380608,
+      "step": 832
+    },
+    {
+      "3d_bbox": 0.755732849240303,
+      "adv": -0.000661049871268915,
+      "bbox": 0.9913836792111397,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.125,
+      "completions/max_terminated_length": 1598.125,
+      "completions/mean_length": 1575.3125,
+      "completions/mean_terminated_length": 1575.3125,
+      "completions/min_length": 1555.625,
+      "completions/min_terminated_length": 1555.625,
+      "entropy": 0.2857117219828069,
+      "epoch": 0.2964008468595625,
+      "geo": 0.8524684980511665,
+      "grad_norm": 0.380859375,
+      "learning_rate": 5.8341708542713575e-06,
+      "loss": 0.0143,
+      "num_tokens": 99213386.0,
+      "point_2d": 0.875,
+      "point_3d": 0.876953125,
+      "reward": 0.8747674003243446,
+      "step": 840
+    },
+    {
+      "3d_bbox": 0.7413047328591347,
+      "adv": 0.0002993446832988411,
+      "bbox": 0.8264872687868774,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.125,
+      "completions/max_terminated_length": 1594.125,
+      "completions/mean_length": 1573.984375,
+      "completions/mean_terminated_length": 1573.984375,
+      "completions/min_length": 1556.75,
+      "completions/min_terminated_length": 1556.75,
+      "entropy": 0.28746917750686407,
+      "epoch": 0.2992237120677488,
+      "geo": 0.9443965330719948,
+      "grad_norm": 0.3359375,
+      "learning_rate": 5.793969849246232e-06,
+      "loss": 0.0195,
+      "num_tokens": 100157950.0,
+      "point_2d": 1.0,
+      "point_3d": 0.9140625,
+      "reward": 0.8704636245965958,
+      "step": 848
+    }
+  ],
+  "logging_steps": 8,
+  "max_steps": 2000,
+  "num_input_tokens_seen": 100393671,
+  "num_train_epochs": 1,
+  "save_steps": 50,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-850/video_preprocessor_config.json

@@ -0,0 +1,44 @@
+{
+  "crop_size": null,
+  "data_format": "channels_first",
+  "default_to_square": true,
+  "device": null,
+  "do_center_crop": null,
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_pad": null,
+  "do_rescale": true,
+  "do_resize": true,
+  "do_sample_frames": false,
+  "fps": null,
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "input_data_format": null,
+  "max_frames": 768,
+  "max_pixels": 12845056,
+  "merge_size": 2,
+  "min_frames": 4,
+  "min_pixels": 3136,
+  "num_frames": null,
+  "patch_size": 14,
+  "processor_class": "Qwen2_5_VLProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "return_metadata": false,
+  "size": {
+    "longest_edge": 12845056,
+    "shortest_edge": 3136
+  },
+  "size_divisor": null,
+  "temporal_patch_size": 2,
+  "video_metadata": null,
+  "video_processor_type": "Qwen2VLVideoProcessor"
+}

checkpoint-950/added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

checkpoint-950/chat_template.jinja

@@ -0,0 +1,7 @@
+{% set image_count = namespace(value=0) %}{% set video_count = namespace(value=0) %}{% for message in messages %}{% if loop.first and message['role'] != 'system' %}<|im_start|>system
+You are a helpful assistant.<|im_end|>
+{% endif %}<|im_start|>{{ message['role'] }}
+{% if message['content'] is string %}{{ message['content'] }}<|im_end|>
+{% else %}{% for content in message['content'] %}{% if content['type'] == 'image' or 'image' in content or 'image_url' in content %}{% set image_count.value = image_count.value + 1 %}{% if add_vision_id %}Picture {{ image_count.value }}: {% endif %}<|vision_start|><|image_pad|><|vision_end|>{% elif content['type'] == 'video' or 'video' in content %}{% set video_count.value = video_count.value + 1 %}{% if add_vision_id %}Video {{ video_count.value }}: {% endif %}<|vision_start|><|video_pad|><|vision_end|>{% elif 'text' in content %}{{ content['text'] }}{% endif %}{% endfor %}<|im_end|>
+{% endif %}{% endfor %}{% if add_generation_prompt %}<|im_start|>assistant
+{% endif %}

checkpoint-950/config.json

@@ -0,0 +1,143 @@
+{
+  "architectures": [
+    "Qwen2_5_PVLForConditionalGeneration"
+  ],
+  "attention_dropout": 0.0,
+  "dtype": "bfloat16",
+  "eos_token_id": 151645,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "image_token_id": 151655,
+  "initializer_range": 0.02,
+  "intermediate_size": 11008,
+  "max_position_embeddings": 128000,
+  "max_window_layers": 70,
+  "model_type": "qwen2_5_vl",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 36,
+  "num_key_value_heads": 2,
+  "pad_token_id": 151643,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": {
+    "mrope_section": [
+      16,
+      24,
+      24
+    ],
+    "rope_type": "default",
+    "type": "default"
+  },
+  "rope_theta": 1000000.0,
+  "sliding_window": 32768,
+  "text_config": {
+    "architectures": [
+      "Qwen2_5_VLForConditionalGeneration"
+    ],
+    "attention_dropout": 0.0,
+    "bos_token_id": 151643,
+    "dtype": "bfloat16",
+    "eos_token_id": 151645,
+    "hidden_act": "silu",
+    "hidden_size": 2048,
+    "image_token_id": null,
+    "initializer_range": 0.02,
+    "intermediate_size": 11008,
+    "layer_types": [
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention",
+      "full_attention"
+    ],
+    "max_position_embeddings": 128000,
+    "max_window_layers": 70,
+    "model_type": "qwen2_5_vl_text",
+    "num_attention_heads": 16,
+    "num_hidden_layers": 36,
+    "num_key_value_heads": 2,
+    "rms_norm_eps": 1e-06,
+    "rope_scaling": {
+      "mrope_section": [
+        16,
+        24,
+        24
+      ],
+      "rope_type": "default",
+      "type": "default"
+    },
+    "rope_theta": 1000000.0,
+    "sliding_window": null,
+    "tie_word_embeddings": true,
+    "use_cache": true,
+    "use_sliding_window": false,
+    "video_token_id": null,
+    "vision_end_token_id": 151653,
+    "vision_start_token_id": 151652,
+    "vision_token_id": 151654,
+    "vocab_size": 151936
+  },
+  "transformers_version": "4.56.2",
+  "use_cache": true,
+  "use_sliding_window": false,
+  "video_token_id": 151656,
+  "vision_config": {
+    "depth": 32,
+    "dtype": "bfloat16",
+    "fullatt_block_indexes": [
+      7,
+      15,
+      23,
+      31
+    ],
+    "hidden_act": "silu",
+    "hidden_size": 1280,
+    "in_channels": 3,
+    "in_chans": 3,
+    "initializer_range": 0.02,
+    "intermediate_size": 3420,
+    "model_type": "qwen2_5_vl",
+    "num_heads": 16,
+    "out_hidden_size": 2048,
+    "patch_size": 14,
+    "spatial_merge_size": 2,
+    "spatial_patch_size": 14,
+    "temporal_patch_size": 2,
+    "tokens_per_second": 2,
+    "window_size": 112
+  },
+  "vision_end_token_id": 151653,
+  "vision_start_token_id": 151652,
+  "vision_token_id": 151654,
+  "vocab_size": 151936
+}

checkpoint-950/generation_config.json

@@ -0,0 +1,11 @@
+{
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "pad_token_id": 151643,
+  "repetition_penalty": 1.05,
+  "temperature": 1e-06,
+  "transformers_version": "4.56.2"
+}

checkpoint-950/preprocessor_config.json

@@ -0,0 +1,37 @@
+{
+  "crop_size": null,
+  "data_format": "channels_first",
+  "default_to_square": true,
+  "device": null,
+  "disable_grouping": null,
+  "do_center_crop": null,
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_processor_type": "Qwen2VLImageProcessorFast",
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "input_data_format": null,
+  "max_pixels": 12845056,
+  "merge_size": 2,
+  "min_pixels": 3136,
+  "patch_size": 14,
+  "processor_class": "Qwen2_5_VLProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "return_tensors": null,
+  "size": {
+    "longest_edge": 12845056,
+    "shortest_edge": 3136
+  },
+  "temporal_patch_size": 2
+}

checkpoint-950/special_tokens_map.json

@@ -0,0 +1,31 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-950/tokenizer_config.json

@@ -0,0 +1,212 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "max_length": null,
+  "model_max_length": 131072,
+  "pad_to_multiple_of": null,
+  "pad_token": "<|endoftext|>",
+  "pad_token_type_id": 0,
+  "padding_side": "left",
+  "processor_class": "Qwen2_5_VLProcessor",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

checkpoint-950/trainer_state.json

@@ -0,0 +1,2748 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.3352152434721242,
+  "eval_steps": 500,
+  "global_step": 950,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "3d_bbox": 0.6610370799899101,
+      "adv": 0.000926484264937244,
+      "bbox": 0.8075725510716438,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.875,
+      "completions/max_terminated_length": 1592.875,
+      "completions/mean_length": 1573.1640625,
+      "completions/mean_terminated_length": 1573.1640625,
+      "completions/min_length": 1554.875,
+      "completions/min_terminated_length": 1554.875,
+      "entropy": 0.2564338077791035,
+      "epoch": 0.0028228652081863093,
+      "geo": 0.7611365988850594,
+      "grad_norm": 0.439453125,
+      "learning_rate": 7e-06,
+      "loss": -0.0002,
+      "num_tokens": 943810.0,
+      "point_2d": 0.875,
+      "point_3d": 0.638671875,
+      "reward": 0.7455703690648079,
+      "step": 8
+    },
+    {
+      "3d_bbox": 0.6985021606087685,
+      "adv": 0.0008333708265126916,
+      "bbox": 0.9951911941170692,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.25,
+      "completions/max_terminated_length": 1594.25,
+      "completions/mean_length": 1573.64453125,
+      "completions/mean_terminated_length": 1573.64453125,
+      "completions/min_length": 1552.375,
+      "completions/min_terminated_length": 1552.375,
+      "entropy": 0.2595143502112478,
+      "epoch": 0.0056457304163726185,
+      "geo": 0.8901385143399239,
+      "grad_norm": 0.341796875,
+      "learning_rate": 9.974874371859297e-06,
+      "loss": 0.0564,
+      "num_tokens": 1888015.0,
+      "point_2d": 1.0,
+      "point_3d": 0.5947265625,
+      "reward": 0.8221049904823303,
+      "step": 16
+    },
+    {
+      "3d_bbox": 0.6706115230917931,
+      "adv": -0.11667590066201683,
+      "bbox": 0.9641365110874176,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1588.75,
+      "completions/max_terminated_length": 1588.75,
+      "completions/mean_length": 1568.69921875,
+      "completions/mean_terminated_length": 1568.69921875,
+      "completions/min_length": 1549.625,
+      "completions/min_terminated_length": 1549.625,
+      "entropy": 0.2620172018650919,
+      "epoch": 0.008468595624558928,
+      "geo": 0.8719085603952408,
+      "grad_norm": 0.31640625,
+      "learning_rate": 9.934673366834172e-06,
+      "loss": 0.0434,
+      "num_tokens": 2830754.0,
+      "point_2d": 0.96484375,
+      "point_3d": 0.7060546875,
+      "reward": 0.8264116197824478,
+      "step": 24
+    },
+    {
+      "3d_bbox": 0.6768141128122807,
+      "adv": 0.00011848201484099263,
+      "bbox": 0.8640391379594803,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.875,
+      "completions/max_terminated_length": 1599.875,
+      "completions/mean_length": 1579.40625,
+      "completions/mean_terminated_length": 1579.40625,
+      "completions/min_length": 1563.0,
+      "completions/min_terminated_length": 1563.0,
+      "entropy": 0.26498199813067913,
+      "epoch": 0.011291460832745237,
+      "geo": 0.8661364912986755,
+      "grad_norm": 0.36328125,
+      "learning_rate": 9.894472361809046e-06,
+      "loss": -0.0005,
+      "num_tokens": 3776674.0,
+      "point_2d": 0.8642578125,
+      "point_3d": 0.603515625,
+      "reward": 0.752156674861908,
+      "step": 32
+    },
+    {
+      "3d_bbox": 0.6714089289307594,
+      "adv": -0.000722882674608627,
+      "bbox": 0.9624817371368408,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1593.75,
+      "completions/max_terminated_length": 1593.75,
+      "completions/mean_length": 1575.52734375,
+      "completions/mean_terminated_length": 1575.52734375,
+      "completions/min_length": 1559.875,
+      "completions/min_terminated_length": 1559.875,
+      "entropy": 0.2738626003265381,
+      "epoch": 0.014114326040931546,
+      "geo": 0.8932898789644241,
+      "grad_norm": 0.341796875,
+      "learning_rate": 9.85427135678392e-06,
+      "loss": 0.0002,
+      "num_tokens": 4721169.0,
+      "point_2d": 0.99609375,
+      "point_3d": 0.6064453125,
+      "reward": 0.8091074377298355,
+      "step": 40
+    },
+    {
+      "3d_bbox": 0.6575617864727974,
+      "adv": -0.0007460190936399158,
+      "bbox": 0.9782034233212471,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.75,
+      "completions/max_terminated_length": 1598.75,
+      "completions/mean_length": 1581.515625,
+      "completions/mean_terminated_length": 1581.515625,
+      "completions/min_length": 1563.875,
+      "completions/min_terminated_length": 1563.875,
+      "entropy": 0.26874606031924486,
+      "epoch": 0.016937191249117856,
+      "geo": 0.922828696668148,
+      "grad_norm": 0.384765625,
+      "learning_rate": 9.814070351758794e-06,
+      "loss": 0.0001,
+      "num_tokens": 5667501.0,
+      "point_2d": 0.9404296875,
+      "point_3d": 0.6171875,
+      "reward": 0.7983455955982208,
+      "step": 48
+    },
+    {
+      "3d_bbox": 0.7516707628965378,
+      "adv": 0.0002944665375252953,
+      "bbox": 0.870270783547312,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1605.5,
+      "completions/max_terminated_length": 1605.5,
+      "completions/mean_length": 1583.109375,
+      "completions/mean_terminated_length": 1583.109375,
+      "completions/min_length": 1563.125,
+      "completions/min_terminated_length": 1563.125,
+      "entropy": 0.27441642340272665,
+      "epoch": 0.019760056457304165,
+      "geo": 0.8707154765725136,
+      "grad_norm": 0.35546875,
+      "learning_rate": 9.773869346733669e-06,
+      "loss": 0.0002,
+      "num_tokens": 6614089.0,
+      "point_2d": 0.875,
+      "point_3d": 0.728515625,
+      "reward": 0.8063642866909504,
+      "step": 56
+    },
+    {
+      "3d_bbox": 0.7704295367002487,
+      "adv": 0.0006564971263287589,
+      "bbox": 0.8711516009643674,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1604.75,
+      "completions/max_terminated_length": 1604.75,
+      "completions/mean_length": 1582.09375,
+      "completions/mean_terminated_length": 1582.09375,
+      "completions/min_length": 1562.25,
+      "completions/min_terminated_length": 1562.25,
+      "entropy": 0.2721484359353781,
+      "epoch": 0.022582921665490474,
+      "geo": 0.8998556286096573,
+      "grad_norm": 0.373046875,
+      "learning_rate": 9.733668341708543e-06,
+      "loss": 0.0156,
+      "num_tokens": 7560513.0,
+      "point_2d": 0.875,
+      "point_3d": 0.7880859375,
+      "reward": 0.8261667564511299,
+      "step": 64
+    },
+    {
+      "3d_bbox": 0.6965150758624077,
+      "adv": 5.7399442994210403e-05,
+      "bbox": 0.9924400746822357,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1603.875,
+      "completions/max_terminated_length": 1603.875,
+      "completions/mean_length": 1580.8203125,
+      "completions/mean_terminated_length": 1580.8203125,
+      "completions/min_length": 1562.0,
+      "completions/min_terminated_length": 1562.0,
+      "entropy": 0.2694994225166738,
+      "epoch": 0.025405786873676783,
+      "geo": 0.7992948293685913,
+      "grad_norm": 0.40625,
+      "learning_rate": 9.693467336683417e-06,
+      "loss": -0.0004,
+      "num_tokens": 8506787.0,
+      "point_2d": 0.875,
+      "point_3d": 0.6201171875,
+      "reward": 0.796018086373806,
+      "step": 72
+    },
+    {
+      "3d_bbox": 0.9022366181015968,
+      "adv": 0.000732647989934776,
+      "bbox": 0.9670402258634567,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.5,
+      "completions/max_terminated_length": 1598.5,
+      "completions/mean_length": 1577.83203125,
+      "completions/mean_terminated_length": 1577.83203125,
+      "completions/min_length": 1560.0,
+      "completions/min_terminated_length": 1560.0,
+      "entropy": 0.2729479926638305,
+      "epoch": 0.028228652081863093,
+      "geo": 0.770210200920701,
+      "grad_norm": 0.28515625,
+      "learning_rate": 9.653266331658291e-06,
+      "loss": 0.0285,
+      "num_tokens": 9452032.0,
+      "point_2d": 0.7578125,
+      "point_3d": 0.8740234375,
+      "reward": 0.8752781972289085,
+      "step": 80
+    },
+    {
+      "3d_bbox": 0.8243749290704727,
+      "adv": 0.0004942354327113208,
+      "bbox": 0.9483200311660767,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.0,
+      "completions/max_terminated_length": 1599.0,
+      "completions/mean_length": 1578.5703125,
+      "completions/mean_terminated_length": 1578.5703125,
+      "completions/min_length": 1564.875,
+      "completions/min_terminated_length": 1564.875,
+      "entropy": 0.27196697471663356,
+      "epoch": 0.031051517290049402,
+      "geo": 0.8520620986819267,
+      "grad_norm": 0.337890625,
+      "learning_rate": 9.613065326633166e-06,
+      "loss": 0.0,
+      "num_tokens": 10397602.0,
+      "point_2d": 0.84375,
+      "point_3d": 0.7431640625,
+      "reward": 0.8399022594094276,
+      "step": 88
+    },
+    {
+      "3d_bbox": 0.7930564805865288,
+      "adv": -0.05789744130743202,
+      "bbox": 0.9452124014496803,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1600.125,
+      "completions/max_terminated_length": 1600.125,
+      "completions/mean_length": 1579.25390625,
+      "completions/mean_terminated_length": 1579.25390625,
+      "completions/min_length": 1560.25,
+      "completions/min_terminated_length": 1560.25,
+      "entropy": 0.2732193539850414,
+      "epoch": 0.03387438249823571,
+      "geo": 0.8466203399002552,
+      "grad_norm": 0.35546875,
+      "learning_rate": 9.57286432160804e-06,
+      "loss": 0.0143,
+      "num_tokens": 11343267.0,
+      "point_2d": 0.7138671875,
+      "point_3d": 0.8203125,
+      "reward": 0.8181121572852135,
+      "step": 96
+    },
+    {
+      "3d_bbox": 0.6107643470168114,
+      "adv": -0.0019174332555849105,
+      "bbox": 0.9693493247032166,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1605.375,
+      "completions/max_terminated_length": 1605.375,
+      "completions/mean_length": 1580.6875,
+      "completions/mean_terminated_length": 1580.6875,
+      "completions/min_length": 1559.75,
+      "completions/min_terminated_length": 1559.75,
+      "entropy": 0.2712484383955598,
+      "epoch": 0.03669724770642202,
+      "geo": 0.760684534907341,
+      "grad_norm": 0.392578125,
+      "learning_rate": 9.532663316582916e-06,
+      "loss": 0.0006,
+      "num_tokens": 12289259.0,
+      "point_2d": 0.849609375,
+      "point_3d": 0.7109375,
+      "reward": 0.7851651459932327,
+      "step": 104
+    },
+    {
+      "3d_bbox": 0.7011105790734291,
+      "adv": 0.000332610031364311,
+      "bbox": 0.997136726975441,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.5,
+      "completions/max_terminated_length": 1601.5,
+      "completions/mean_length": 1578.38671875,
+      "completions/mean_terminated_length": 1578.38671875,
+      "completions/min_length": 1560.0,
+      "completions/min_terminated_length": 1560.0,
+      "entropy": 0.2741631781682372,
+      "epoch": 0.03952011291460833,
+      "geo": 0.868082121014595,
+      "grad_norm": 0.349609375,
+      "learning_rate": 9.49246231155779e-06,
+      "loss": -0.0005,
+      "num_tokens": 13234494.0,
+      "point_2d": 0.875,
+      "point_3d": 0.759765625,
+      "reward": 0.8332532122731209,
+      "step": 112
+    },
+    {
+      "3d_bbox": 0.6414853371679783,
+      "adv": 6.49127487122314e-05,
+      "bbox": 0.9946233779191971,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.875,
+      "completions/max_terminated_length": 1599.875,
+      "completions/mean_length": 1580.48828125,
+      "completions/mean_terminated_length": 1580.48828125,
+      "completions/min_length": 1562.5,
+      "completions/min_terminated_length": 1562.5,
+      "entropy": 0.2765682805329561,
+      "epoch": 0.04234297812279464,
+      "geo": 0.7812720276415348,
+      "grad_norm": 0.392578125,
+      "learning_rate": 9.452261306532664e-06,
+      "loss": -0.0002,
+      "num_tokens": 14180739.0,
+      "point_2d": 0.875,
+      "point_3d": 0.7275390625,
+      "reward": 0.8096619322896004,
+      "step": 120
+    },
+    {
+      "3d_bbox": 0.696478821337223,
+      "adv": 0.00018893864194069465,
+      "bbox": 0.9915411844849586,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1600.75,
+      "completions/max_terminated_length": 1600.75,
+      "completions/mean_length": 1581.48828125,
+      "completions/mean_terminated_length": 1581.48828125,
+      "completions/min_length": 1566.25,
+      "completions/min_terminated_length": 1566.25,
+      "entropy": 0.27939224475994706,
+      "epoch": 0.04516584333098095,
+      "geo": 0.9225678443908691,
+      "grad_norm": 0.359375,
+      "learning_rate": 9.412060301507539e-06,
+      "loss": 0.0004,
+      "num_tokens": 15126808.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8037109375,
+      "reward": 0.8729327321052551,
+      "step": 128
+    },
+    {
+      "3d_bbox": 0.7811713069677353,
+      "adv": -0.00017097587488024146,
+      "bbox": 0.983598880469799,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.0,
+      "completions/max_terminated_length": 1599.0,
+      "completions/mean_length": 1575.9765625,
+      "completions/mean_terminated_length": 1575.9765625,
+      "completions/min_length": 1560.875,
+      "completions/min_terminated_length": 1560.875,
+      "entropy": 0.2712345658801496,
+      "epoch": 0.04798870853916726,
+      "geo": 0.829373586922884,
+      "grad_norm": 0.365234375,
+      "learning_rate": 9.371859296482413e-06,
+      "loss": 0.0002,
+      "num_tokens": 16071682.0,
+      "point_2d": 0.75,
+      "point_3d": 0.76171875,
+      "reward": 0.819122239947319,
+      "step": 136
+    },
+    {
+      "3d_bbox": 0.6566238105297089,
+      "adv": 5.832723485355018e-05,
+      "bbox": 0.9859664067626,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1597.5,
+      "completions/max_terminated_length": 1597.5,
+      "completions/mean_length": 1580.12890625,
+      "completions/mean_terminated_length": 1580.12890625,
+      "completions/min_length": 1565.0,
+      "completions/min_terminated_length": 1565.0,
+      "entropy": 0.2754409732297063,
+      "epoch": 0.05081157374735357,
+      "geo": 0.8895541727542877,
+      "grad_norm": 0.318359375,
+      "learning_rate": 9.331658291457287e-06,
+      "loss": 0.0218,
+      "num_tokens": 17017499.0,
+      "point_2d": 1.0,
+      "point_3d": 0.6328125,
+      "reward": 0.8188506811857224,
+      "step": 144
+    },
+    {
+      "3d_bbox": 0.655942264944315,
+      "adv": 0.00024080171715468168,
+      "bbox": 0.9830196872353554,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1600.125,
+      "completions/max_terminated_length": 1600.125,
+      "completions/mean_length": 1581.3828125,
+      "completions/mean_terminated_length": 1581.3828125,
+      "completions/min_length": 1564.0,
+      "completions/min_terminated_length": 1564.0,
+      "entropy": 0.2764031286351383,
+      "epoch": 0.053634438955539876,
+      "geo": 0.8772727139294147,
+      "grad_norm": 0.294921875,
+      "learning_rate": 9.291457286432161e-06,
+      "loss": 0.015,
+      "num_tokens": 17963757.0,
+      "point_2d": 0.875,
+      "point_3d": 0.7880859375,
+      "reward": 0.8255119547247887,
+      "step": 152
+    },
+    {
+      "3d_bbox": 0.8087940737605095,
+      "adv": -0.062093503400319605,
+      "bbox": 0.9799230694770813,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.5,
+      "completions/max_terminated_length": 1598.5,
+      "completions/mean_length": 1579.2421875,
+      "completions/mean_terminated_length": 1579.2421875,
+      "completions/min_length": 1560.25,
+      "completions/min_terminated_length": 1560.25,
+      "entropy": 0.27893486013635993,
+      "epoch": 0.056457304163726185,
+      "geo": 0.8533996716141701,
+      "grad_norm": 0.314453125,
+      "learning_rate": 9.251256281407036e-06,
+      "loss": 0.0157,
+      "num_tokens": 18909307.0,
+      "point_2d": 0.859375,
+      "point_3d": 0.814453125,
+      "reward": 0.8656363189220428,
+      "step": 160
+    },
+    {
+      "3d_bbox": 0.7417622283101082,
+      "adv": 0.00038278038937278325,
+      "bbox": 0.9947961643338203,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.875,
+      "completions/max_terminated_length": 1596.875,
+      "completions/mean_length": 1578.4765625,
+      "completions/mean_terminated_length": 1578.4765625,
+      "completions/min_length": 1562.375,
+      "completions/min_terminated_length": 1562.375,
+      "entropy": 0.2772809020243585,
+      "epoch": 0.059280169371912494,
+      "geo": 0.8446554020047188,
+      "grad_norm": 0.337890625,
+      "learning_rate": 9.21105527638191e-06,
+      "loss": -0.0001,
+      "num_tokens": 19854893.0,
+      "point_2d": 0.875,
+      "point_3d": 0.7666015625,
+      "reward": 0.844539999961853,
+      "step": 168
+    },
+    {
+      "3d_bbox": 0.7199448123574257,
+      "adv": -0.00025279237161157653,
+      "bbox": 0.9939833134412766,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.375,
+      "completions/max_terminated_length": 1599.375,
+      "completions/mean_length": 1577.875,
+      "completions/mean_terminated_length": 1577.875,
+      "completions/min_length": 1560.875,
+      "completions/min_terminated_length": 1560.875,
+      "entropy": 0.2748410953208804,
+      "epoch": 0.062103034580098804,
+      "geo": 0.8426482602953911,
+      "grad_norm": 0.353515625,
+      "learning_rate": 9.170854271356784e-06,
+      "loss": 0.0001,
+      "num_tokens": 20800093.0,
+      "point_2d": 0.8935546875,
+      "point_3d": 0.7666015625,
+      "reward": 0.8435210958123207,
+      "step": 176
+    },
+    {
+      "3d_bbox": 0.8168098330497742,
+      "adv": 0.0003567840885807527,
+      "bbox": 0.8971930183470249,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.5,
+      "completions/max_terminated_length": 1601.5,
+      "completions/mean_length": 1581.01171875,
+      "completions/mean_terminated_length": 1581.01171875,
+      "completions/min_length": 1562.625,
+      "completions/min_terminated_length": 1562.625,
+      "entropy": 0.27716846112161875,
+      "epoch": 0.06492589978828511,
+      "geo": 0.8458116948604584,
+      "grad_norm": 0.376953125,
+      "learning_rate": 9.13065326633166e-06,
+      "loss": -0.0,
+      "num_tokens": 21746232.0,
+      "point_2d": 0.736328125,
+      "point_3d": 0.859375,
+      "reward": 0.8274264857172966,
+      "step": 184
+    },
+    {
+      "3d_bbox": 0.7658436857163906,
+      "adv": -0.00043575108065851964,
+      "bbox": 0.994478665292263,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.5,
+      "completions/max_terminated_length": 1599.5,
+      "completions/mean_length": 1577.2265625,
+      "completions/mean_terminated_length": 1577.2265625,
+      "completions/min_length": 1558.25,
+      "completions/min_terminated_length": 1558.25,
+      "entropy": 0.28411462996155024,
+      "epoch": 0.06774876499647142,
+      "geo": 0.8858354687690735,
+      "grad_norm": 0.404296875,
+      "learning_rate": 9.090452261306534e-06,
+      "loss": -0.0,
+      "num_tokens": 22691162.0,
+      "point_2d": 1.0,
+      "point_3d": 0.83203125,
+      "reward": 0.8980883881449699,
+      "step": 192
+    },
+    {
+      "3d_bbox": 0.689879797399044,
+      "adv": 0.0008139211470279406,
+      "bbox": 0.8756355233490467,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.25,
+      "completions/max_terminated_length": 1595.25,
+      "completions/mean_length": 1579.328125,
+      "completions/mean_terminated_length": 1579.328125,
+      "completions/min_length": 1562.125,
+      "completions/min_terminated_length": 1562.125,
+      "entropy": 0.27792059257626534,
+      "epoch": 0.07057163020465773,
+      "geo": 0.8776071295142174,
+      "grad_norm": 0.28125,
+      "learning_rate": 9.050251256281409e-06,
+      "loss": -0.0006,
+      "num_tokens": 23637118.0,
+      "point_2d": 0.875,
+      "point_3d": 0.765625,
+      "reward": 0.8015350699424744,
+      "step": 200
+    },
+    {
+      "3d_bbox": 0.7628588750958443,
+      "adv": 0.0005178588535272866,
+      "bbox": 0.9862369671463966,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1603.5,
+      "completions/max_terminated_length": 1603.5,
+      "completions/mean_length": 1582.5,
+      "completions/mean_terminated_length": 1582.5,
+      "completions/min_length": 1561.75,
+      "completions/min_terminated_length": 1561.75,
+      "entropy": 0.2777437614277005,
+      "epoch": 0.07339449541284404,
+      "geo": 0.7896297052502632,
+      "grad_norm": 0.318359375,
+      "learning_rate": 9.010050251256283e-06,
+      "loss": -0.0003,
+      "num_tokens": 24583790.0,
+      "point_2d": 0.75,
+      "point_3d": 0.822265625,
+      "reward": 0.8303403630852699,
+      "step": 208
+    },
+    {
+      "3d_bbox": 0.6901299897581339,
+      "adv": 1.327232166659087e-05,
+      "bbox": 0.9742533341050148,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.375,
+      "completions/max_terminated_length": 1595.375,
+      "completions/mean_length": 1580.3828125,
+      "completions/mean_terminated_length": 1580.3828125,
+      "completions/min_length": 1566.125,
+      "completions/min_terminated_length": 1566.125,
+      "entropy": 0.2789337821304798,
+      "epoch": 0.07621736062103035,
+      "geo": 0.9093400537967682,
+      "grad_norm": 0.33984375,
+      "learning_rate": 8.969849246231157e-06,
+      "loss": 0.0001,
+      "num_tokens": 25529904.0,
+      "point_2d": 1.0,
+      "point_3d": 0.724609375,
+      "reward": 0.8472481593489647,
+      "step": 216
+    },
+    {
+      "3d_bbox": 0.8317120671272278,
+      "adv": 0.0007813970478309784,
+      "bbox": 0.99237409979105,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.375,
+      "completions/max_terminated_length": 1592.375,
+      "completions/mean_length": 1576.484375,
+      "completions/mean_terminated_length": 1576.484375,
+      "completions/min_length": 1559.125,
+      "completions/min_terminated_length": 1559.125,
+      "entropy": 0.281503742095083,
+      "epoch": 0.07904022582921666,
+      "geo": 0.9148535653948784,
+      "grad_norm": 0.33203125,
+      "learning_rate": 8.92964824120603e-06,
+      "loss": -0.0005,
+      "num_tokens": 26474884.0,
+      "point_2d": 1.0,
+      "point_3d": 0.859375,
+      "reward": 0.9208653047680855,
+      "step": 224
+    },
+    {
+      "3d_bbox": 0.8539343103766441,
+      "adv": 0.0006083497164581786,
+      "bbox": 0.9947174340486526,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1607.25,
+      "completions/max_terminated_length": 1607.25,
+      "completions/mean_length": 1582.21875,
+      "completions/mean_terminated_length": 1582.21875,
+      "completions/min_length": 1562.75,
+      "completions/min_terminated_length": 1562.75,
+      "entropy": 0.28075447864830494,
+      "epoch": 0.08186309103740297,
+      "geo": 0.9205377101898193,
+      "grad_norm": 0.3828125,
+      "learning_rate": 8.889447236180904e-06,
+      "loss": 0.0151,
+      "num_tokens": 27421356.0,
+      "point_2d": 0.9423828125,
+      "point_3d": 0.919921875,
+      "reward": 0.9277391210198402,
+      "step": 232
+    },
+    {
+      "3d_bbox": 0.8143627047538757,
+      "adv": 0.0002584877729532309,
+      "bbox": 0.9961168393492699,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1604.375,
+      "completions/max_terminated_length": 1604.375,
+      "completions/mean_length": 1581.84375,
+      "completions/mean_terminated_length": 1581.84375,
+      "completions/min_length": 1562.0,
+      "completions/min_terminated_length": 1562.0,
+      "entropy": 0.2808487247675657,
+      "epoch": 0.08468595624558928,
+      "geo": 0.9482192173600197,
+      "grad_norm": 0.3125,
+      "learning_rate": 8.84924623115578e-06,
+      "loss": 0.015,
+      "num_tokens": 28367588.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8505859375,
+      "reward": 0.9152663722634315,
+      "step": 240
+    },
+    {
+      "3d_bbox": 0.7562066540122032,
+      "adv": 0.0011895421193912625,
+      "bbox": 0.9870459213852882,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1605.625,
+      "completions/max_terminated_length": 1605.625,
+      "completions/mean_length": 1582.22265625,
+      "completions/mean_terminated_length": 1582.22265625,
+      "completions/min_length": 1560.125,
+      "completions/min_terminated_length": 1560.125,
+      "entropy": 0.2828774293884635,
+      "epoch": 0.08750882145377559,
+      "geo": 0.8906307443976402,
+      "grad_norm": 0.310546875,
+      "learning_rate": 8.809045226130654e-06,
+      "loss": -0.0001,
+      "num_tokens": 29314053.0,
+      "point_2d": 1.0,
+      "point_3d": 0.84765625,
+      "reward": 0.8977271988987923,
+      "step": 248
+    },
+    {
+      "3d_bbox": 0.662892609834671,
+      "adv": -0.0004179688154977157,
+      "bbox": 0.9923923686146736,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.875,
+      "completions/max_terminated_length": 1598.875,
+      "completions/mean_length": 1579.796875,
+      "completions/mean_terminated_length": 1579.796875,
+      "completions/min_length": 1563.125,
+      "completions/min_terminated_length": 1563.125,
+      "entropy": 0.27959624165669084,
+      "epoch": 0.0903316866619619,
+      "geo": 0.9055434167385101,
+      "grad_norm": 0.259765625,
+      "learning_rate": 8.768844221105528e-06,
+      "loss": -0.0002,
+      "num_tokens": 30259921.0,
+      "point_2d": 0.7509765625,
+      "point_3d": 0.8154296875,
+      "reward": 0.805422805249691,
+      "step": 256
+    },
+    {
+      "3d_bbox": 0.7727803811430931,
+      "adv": -0.0010380800276834634,
+      "bbox": 0.9807735681533813,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1604.375,
+      "completions/max_terminated_length": 1604.375,
+      "completions/mean_length": 1581.1796875,
+      "completions/mean_terminated_length": 1581.1796875,
+      "completions/min_length": 1564.375,
+      "completions/min_terminated_length": 1564.375,
+      "entropy": 0.28243178501725197,
+      "epoch": 0.0931545518701482,
+      "geo": 0.783092018100433,
+      "grad_norm": 0.3203125,
+      "learning_rate": 8.728643216080403e-06,
+      "loss": 0.0003,
+      "num_tokens": 31205887.0,
+      "point_2d": 0.75,
+      "point_3d": 0.8544921875,
+      "reward": 0.8395115286111832,
+      "step": 264
+    },
+    {
+      "3d_bbox": 0.7379852011799812,
+      "adv": -0.0005326499958755448,
+      "bbox": 0.9934581220149994,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.875,
+      "completions/max_terminated_length": 1598.875,
+      "completions/mean_length": 1579.44921875,
+      "completions/mean_terminated_length": 1579.44921875,
+      "completions/min_length": 1560.375,
+      "completions/min_terminated_length": 1560.375,
+      "entropy": 0.2767978576011956,
+      "epoch": 0.09597741707833451,
+      "geo": 0.9199443385004997,
+      "grad_norm": 0.40625,
+      "learning_rate": 8.688442211055277e-06,
+      "loss": 0.0002,
+      "num_tokens": 32151650.0,
+      "point_2d": 0.9990234375,
+      "point_3d": 0.8896484375,
+      "reward": 0.9050287976861,
+      "step": 272
+    },
+    {
+      "3d_bbox": 0.6362542062997818,
+      "adv": 0.0003683161558001302,
+      "bbox": 0.9273895137012005,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1603.375,
+      "completions/max_terminated_length": 1603.375,
+      "completions/mean_length": 1583.1875,
+      "completions/mean_terminated_length": 1583.1875,
+      "completions/min_length": 1566.25,
+      "completions/min_terminated_length": 1566.25,
+      "entropy": 0.28121166303753853,
+      "epoch": 0.09880028228652082,
+      "geo": 0.6698016822338104,
+      "grad_norm": 0.419921875,
+      "learning_rate": 8.648241206030151e-06,
+      "loss": 0.0104,
+      "num_tokens": 33098514.0,
+      "point_2d": 0.748046875,
+      "point_3d": 0.8564453125,
+      "reward": 0.7920339703559875,
+      "step": 280
+    },
+    {
+      "3d_bbox": 0.8045329824090004,
+      "adv": 0.0005392557122831931,
+      "bbox": 0.9957676902413368,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.75,
+      "completions/max_terminated_length": 1598.75,
+      "completions/mean_length": 1582.8515625,
+      "completions/mean_terminated_length": 1582.8515625,
+      "completions/min_length": 1564.75,
+      "completions/min_terminated_length": 1564.75,
+      "entropy": 0.2840573275461793,
+      "epoch": 0.10162314749470713,
+      "geo": 0.894830085337162,
+      "grad_norm": 0.263671875,
+      "learning_rate": 8.608040201005025e-06,
+      "loss": -0.0003,
+      "num_tokens": 34045156.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8974609375,
+      "reward": 0.9244403913617134,
+      "step": 288
+    },
+    {
+      "3d_bbox": 0.6487490832805634,
+      "adv": -0.036328162997961044,
+      "bbox": 0.9537674784660339,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1606.125,
+      "completions/max_terminated_length": 1606.125,
+      "completions/mean_length": 1583.43359375,
+      "completions/mean_terminated_length": 1583.43359375,
+      "completions/min_length": 1564.25,
+      "completions/min_terminated_length": 1564.25,
+      "entropy": 0.2789871282875538,
+      "epoch": 0.10444601270289344,
+      "geo": 0.8219972774386406,
+      "grad_norm": 0.3671875,
+      "learning_rate": 8.5678391959799e-06,
+      "loss": 0.0091,
+      "num_tokens": 34991979.0,
+      "point_2d": 0.875,
+      "point_3d": 0.8125,
+      "reward": 0.8225041329860687,
+      "step": 296
+    },
+    {
+      "3d_bbox": 0.7852863818407059,
+      "adv": -2.922765997936949e-05,
+      "bbox": 0.9355472847819328,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.5,
+      "completions/max_terminated_length": 1601.5,
+      "completions/mean_length": 1579.7890625,
+      "completions/mean_terminated_length": 1579.7890625,
+      "completions/min_length": 1561.5,
+      "completions/min_terminated_length": 1561.5,
+      "entropy": 0.2808959139510989,
+      "epoch": 0.10726887791107975,
+      "geo": 0.7100357487797737,
+      "grad_norm": 0.333984375,
+      "learning_rate": 8.527638190954774e-06,
+      "loss": 0.0003,
+      "num_tokens": 35937853.0,
+      "point_2d": 0.6728515625,
+      "point_3d": 0.798828125,
+      "reward": 0.7981283441185951,
+      "step": 304
+    },
+    {
+      "3d_bbox": 0.6897341758012772,
+      "adv": -0.0010775219143397408,
+      "bbox": 0.9957167953252792,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.0,
+      "completions/max_terminated_length": 1598.0,
+      "completions/mean_length": 1578.40234375,
+      "completions/mean_terminated_length": 1578.40234375,
+      "completions/min_length": 1560.625,
+      "completions/min_terminated_length": 1560.625,
+      "entropy": 0.28013237891718745,
+      "epoch": 0.11009174311926606,
+      "geo": 0.8601852916181087,
+      "grad_norm": 0.322265625,
+      "learning_rate": 8.487437185929648e-06,
+      "loss": 0.0003,
+      "num_tokens": 36883412.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8154296875,
+      "reward": 0.8752201721072197,
+      "step": 312
+    },
+    {
+      "3d_bbox": 0.7732156217098236,
+      "adv": 0.00015621728016412817,
+      "bbox": 0.9900884553790092,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.25,
+      "completions/max_terminated_length": 1595.25,
+      "completions/mean_length": 1574.06640625,
+      "completions/mean_terminated_length": 1574.06640625,
+      "completions/min_length": 1556.5,
+      "completions/min_terminated_length": 1556.5,
+      "entropy": 0.2842056662775576,
+      "epoch": 0.11291460832745237,
+      "geo": 0.8583031129091978,
+      "grad_norm": 0.359375,
+      "learning_rate": 8.447236180904524e-06,
+      "loss": -0.0003,
+      "num_tokens": 37827701.0,
+      "point_2d": 0.8583984375,
+      "point_3d": 0.8681640625,
+      "reward": 0.8724666386842728,
+      "step": 320
+    },
+    {
+      "3d_bbox": 0.7427574843168259,
+      "adv": -0.00022968700068304315,
+      "bbox": 0.9458357989788055,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.25,
+      "completions/max_terminated_length": 1599.25,
+      "completions/mean_length": 1576.78125,
+      "completions/mean_terminated_length": 1576.78125,
+      "completions/min_length": 1555.625,
+      "completions/min_terminated_length": 1555.625,
+      "entropy": 0.28632913110777736,
+      "epoch": 0.11573747353563868,
+      "geo": 0.9527824521064758,
+      "grad_norm": 0.2578125,
+      "learning_rate": 8.407035175879398e-06,
+      "loss": -0.0,
+      "num_tokens": 38772853.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8623046875,
+      "reward": 0.8877245113253593,
+      "step": 328
+    },
+    {
+      "3d_bbox": 0.7519282810389996,
+      "adv": 0.0003111484527380526,
+      "bbox": 0.9336327686905861,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.0,
+      "completions/max_terminated_length": 1598.0,
+      "completions/mean_length": 1578.54296875,
+      "completions/mean_terminated_length": 1578.54296875,
+      "completions/min_length": 1560.0,
+      "completions/min_terminated_length": 1560.0,
+      "entropy": 0.2819040766917169,
+      "epoch": 0.11856033874382499,
+      "geo": 0.9259950742125511,
+      "grad_norm": 0.306640625,
+      "learning_rate": 8.366834170854273e-06,
+      "loss": 0.0202,
+      "num_tokens": 39718400.0,
+      "point_2d": 0.7451171875,
+      "point_3d": 0.748046875,
+      "reward": 0.7946812808513641,
+      "step": 336
+    },
+    {
+      "3d_bbox": 0.6988782063126564,
+      "adv": 8.153711894226534e-05,
+      "bbox": 0.9918540865182877,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1602.0,
+      "completions/max_terminated_length": 1602.0,
+      "completions/mean_length": 1578.3984375,
+      "completions/mean_terminated_length": 1578.3984375,
+      "completions/min_length": 1557.0,
+      "completions/min_terminated_length": 1557.0,
+      "entropy": 0.2852953290566802,
+      "epoch": 0.1213832039520113,
+      "geo": 0.824953593313694,
+      "grad_norm": 0.33984375,
+      "learning_rate": 8.326633165829147e-06,
+      "loss": 0.0003,
+      "num_tokens": 40663774.0,
+      "point_2d": 0.875,
+      "point_3d": 0.771484375,
+      "reward": 0.8343041799962521,
+      "step": 344
+    },
+    {
+      "3d_bbox": 0.7085578441619873,
+      "adv": 0.000511349891894497,
+      "bbox": 0.9954689294099808,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.625,
+      "completions/max_terminated_length": 1601.625,
+      "completions/mean_length": 1578.21484375,
+      "completions/mean_terminated_length": 1578.21484375,
+      "completions/min_length": 1557.5,
+      "completions/min_terminated_length": 1557.5,
+      "entropy": 0.2835490182042122,
+      "epoch": 0.12420606916019761,
+      "geo": 0.8287867829203606,
+      "grad_norm": 0.31640625,
+      "learning_rate": 8.286432160804021e-06,
+      "loss": 0.0,
+      "num_tokens": 41609261.0,
+      "point_2d": 0.875,
+      "point_3d": 0.8564453125,
+      "reward": 0.8588679879903793,
+      "step": 352
+    },
+    {
+      "3d_bbox": 0.7324622049927711,
+      "adv": -0.06052318853471661,
+      "bbox": 0.9785870239138603,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.75,
+      "completions/max_terminated_length": 1599.75,
+      "completions/mean_length": 1580.8671875,
+      "completions/mean_terminated_length": 1580.8671875,
+      "completions/min_length": 1562.875,
+      "completions/min_terminated_length": 1562.875,
+      "entropy": 0.2788377907127142,
+      "epoch": 0.12702893436838392,
+      "geo": 0.8886202499270439,
+      "grad_norm": 0.765625,
+      "learning_rate": 8.246231155778895e-06,
+      "loss": 0.0295,
+      "num_tokens": 42555355.0,
+      "point_2d": 0.984375,
+      "point_3d": 0.80078125,
+      "reward": 0.8740513622760773,
+      "step": 360
+    },
+    {
+      "3d_bbox": 0.6365351751446724,
+      "adv": 0.0003275554918218404,
+      "bbox": 0.8720735665410757,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.875,
+      "completions/max_terminated_length": 1601.875,
+      "completions/mean_length": 1577.73046875,
+      "completions/mean_terminated_length": 1577.73046875,
+      "completions/min_length": 1555.75,
+      "completions/min_terminated_length": 1555.75,
+      "entropy": 0.28446380188688636,
+      "epoch": 0.12985179957657023,
+      "geo": 0.8144352469826117,
+      "grad_norm": 0.34375,
+      "learning_rate": 8.20603015075377e-06,
+      "loss": -0.0001,
+      "num_tokens": 43500534.0,
+      "point_2d": 0.875,
+      "point_3d": 0.7666015625,
+      "reward": 0.787552572786808,
+      "step": 368
+    },
+    {
+      "3d_bbox": 0.7198036424815655,
+      "adv": -9.503715409664437e-05,
+      "bbox": 0.9939679279923439,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.75,
+      "completions/max_terminated_length": 1592.75,
+      "completions/mean_length": 1573.19140625,
+      "completions/mean_terminated_length": 1573.19140625,
+      "completions/min_length": 1557.0,
+      "completions/min_terminated_length": 1557.0,
+      "entropy": 0.28550330083817244,
+      "epoch": 0.13267466478475654,
+      "geo": 0.9650512784719467,
+      "grad_norm": 0.359375,
+      "learning_rate": 8.165829145728644e-06,
+      "loss": 0.0004,
+      "num_tokens": 44444583.0,
+      "point_2d": 0.9892578125,
+      "point_3d": 0.83984375,
+      "reward": 0.8857182934880257,
+      "step": 376
+    },
+    {
+      "3d_bbox": 0.7569727748632431,
+      "adv": 0.0004161505894444417,
+      "bbox": 0.9042535312473774,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.5,
+      "completions/max_terminated_length": 1599.5,
+      "completions/mean_length": 1577.55859375,
+      "completions/mean_terminated_length": 1577.55859375,
+      "completions/min_length": 1561.375,
+      "completions/min_terminated_length": 1561.375,
+      "entropy": 0.2907078079879284,
+      "epoch": 0.13549752999294284,
+      "geo": 0.6764855124056339,
+      "grad_norm": 0.345703125,
+      "learning_rate": 8.125628140703518e-06,
+      "loss": 0.0001,
+      "num_tokens": 45389846.0,
+      "point_2d": 0.69140625,
+      "point_3d": 0.8193359375,
+      "reward": 0.7929921373724937,
+      "step": 384
+    },
+    {
+      "3d_bbox": 0.6764840334653854,
+      "adv": -1.4531655324390158e-05,
+      "bbox": 0.8646975085139275,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1600.5,
+      "completions/max_terminated_length": 1600.5,
+      "completions/mean_length": 1579.08984375,
+      "completions/mean_terminated_length": 1579.08984375,
+      "completions/min_length": 1558.375,
+      "completions/min_terminated_length": 1558.375,
+      "entropy": 0.2863780874758959,
+      "epoch": 0.13832039520112915,
+      "geo": 0.5404472178779542,
+      "grad_norm": 0.3125,
+      "learning_rate": 8.085427135678392e-06,
+      "loss": 0.0,
+      "num_tokens": 46335629.0,
+      "point_2d": 0.5048828125,
+      "point_3d": 0.806640625,
+      "reward": 0.713176254183054,
+      "step": 392
+    },
+    {
+      "3d_bbox": 0.7255214750766754,
+      "adv": -0.0006330321593850385,
+      "bbox": 0.891293590888381,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1602.375,
+      "completions/max_terminated_length": 1602.375,
+      "completions/mean_length": 1576.9921875,
+      "completions/mean_terminated_length": 1576.9921875,
+      "completions/min_length": 1558.125,
+      "completions/min_terminated_length": 1558.125,
+      "entropy": 0.28716066014021635,
+      "epoch": 0.14114326040931546,
+      "geo": 0.9741705134510994,
+      "grad_norm": 0.314453125,
+      "learning_rate": 8.045226130653267e-06,
+      "loss": 0.0004,
+      "num_tokens": 47280507.0,
+      "point_2d": 0.8916015625,
+      "point_3d": 0.859375,
+      "reward": 0.8419479243457317,
+      "step": 400
+    },
+    {
+      "3d_bbox": 0.7029357459396124,
+      "adv": 0.0008838025460136123,
+      "bbox": 0.9950641989707947,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.125,
+      "completions/max_terminated_length": 1598.125,
+      "completions/mean_length": 1575.92578125,
+      "completions/mean_terminated_length": 1575.92578125,
+      "completions/min_length": 1557.75,
+      "completions/min_terminated_length": 1557.75,
+      "entropy": 0.2837083409540355,
+      "epoch": 0.14396612561750177,
+      "geo": 0.864502564072609,
+      "grad_norm": 0.345703125,
+      "learning_rate": 8.005025125628141e-06,
+      "loss": -0.0003,
+      "num_tokens": 48225240.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8310546875,
+      "reward": 0.8822636753320694,
+      "step": 408
+    },
+    {
+      "3d_bbox": 0.6913789249956608,
+      "adv": -0.0002196824352722615,
+      "bbox": 0.9765095412731171,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.125,
+      "completions/max_terminated_length": 1598.125,
+      "completions/mean_length": 1579.28515625,
+      "completions/mean_terminated_length": 1579.28515625,
+      "completions/min_length": 1559.25,
+      "completions/min_terminated_length": 1559.25,
+      "entropy": 0.28565866220742464,
+      "epoch": 0.14678899082568808,
+      "geo": 0.8599895760416985,
+      "grad_norm": 0.30859375,
+      "learning_rate": 7.964824120603015e-06,
+      "loss": 0.0001,
+      "num_tokens": 49171057.0,
+      "point_2d": 1.0,
+      "point_3d": 0.904296875,
+      "reward": 0.8930463194847107,
+      "step": 416
+    },
+    {
+      "3d_bbox": 0.6788053698837757,
+      "adv": -0.00020617676273104735,
+      "bbox": 0.9748031497001648,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.25,
+      "completions/max_terminated_length": 1595.25,
+      "completions/mean_length": 1577.51953125,
+      "completions/mean_terminated_length": 1577.51953125,
+      "completions/min_length": 1559.625,
+      "completions/min_terminated_length": 1559.625,
+      "entropy": 0.28440852370113134,
+      "epoch": 0.1496118560338744,
+      "geo": 0.822394497692585,
+      "grad_norm": 0.40234375,
+      "learning_rate": 7.92462311557789e-06,
+      "loss": 0.0006,
+      "num_tokens": 50116302.0,
+      "point_2d": 0.875,
+      "point_3d": 0.8310546875,
+      "reward": 0.8399158045649529,
+      "step": 424
+    },
+    {
+      "3d_bbox": 0.6794650703668594,
+      "adv": 0.0011559298304746335,
+      "bbox": 0.9097459949553013,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.625,
+      "completions/max_terminated_length": 1596.625,
+      "completions/mean_length": 1573.80859375,
+      "completions/mean_terminated_length": 1573.80859375,
+      "completions/min_length": 1552.625,
+      "completions/min_terminated_length": 1552.625,
+      "entropy": 0.28376144589856267,
+      "epoch": 0.1524347212420607,
+      "geo": 0.7893084269016981,
+      "grad_norm": 0.3203125,
+      "learning_rate": 7.884422110552764e-06,
+      "loss": 0.0,
+      "num_tokens": 51060365.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8681640625,
+      "reward": 0.8643437847495079,
+      "step": 432
+    },
+    {
+      "3d_bbox": 0.7942164540290833,
+      "adv": -3.183347871527076e-05,
+      "bbox": 0.9926317036151886,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.75,
+      "completions/max_terminated_length": 1598.75,
+      "completions/mean_length": 1577.796875,
+      "completions/mean_terminated_length": 1577.796875,
+      "completions/min_length": 1553.125,
+      "completions/min_terminated_length": 1553.125,
+      "entropy": 0.2821194725111127,
+      "epoch": 0.155257586450247,
+      "geo": 0.9503362327814102,
+      "grad_norm": 0.3984375,
+      "learning_rate": 7.844221105527638e-06,
+      "loss": 0.0002,
+      "num_tokens": 52005633.0,
+      "point_2d": 1.0,
+      "point_3d": 0.869140625,
+      "reward": 0.9139972031116486,
+      "step": 440
+    },
+    {
+      "3d_bbox": 0.6752692721784115,
+      "adv": -9.695853077573702e-05,
+      "bbox": 0.982423759996891,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.25,
+      "completions/max_terminated_length": 1596.25,
+      "completions/mean_length": 1574.5546875,
+      "completions/mean_terminated_length": 1574.5546875,
+      "completions/min_length": 1555.125,
+      "completions/min_terminated_length": 1555.125,
+      "entropy": 0.2852726257406175,
+      "epoch": 0.15808045165843332,
+      "geo": 0.5512092066928744,
+      "grad_norm": 0.322265625,
+      "learning_rate": 7.804020100502512e-06,
+      "loss": -0.0004,
+      "num_tokens": 52949951.0,
+      "point_2d": 0.75,
+      "point_3d": 0.7998046875,
+      "reward": 0.801874428987503,
+      "step": 448
+    },
+    {
+      "3d_bbox": 0.7081510461866856,
+      "adv": 0.0003901398231391795,
+      "bbox": 0.9892760217189789,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.875,
+      "completions/max_terminated_length": 1596.875,
+      "completions/mean_length": 1573.0390625,
+      "completions/mean_terminated_length": 1573.0390625,
+      "completions/min_length": 1551.875,
+      "completions/min_terminated_length": 1551.875,
+      "entropy": 0.2837893716059625,
+      "epoch": 0.16090331686661963,
+      "geo": 0.9603376165032387,
+      "grad_norm": 0.3125,
+      "learning_rate": 7.763819095477387e-06,
+      "loss": -0.0007,
+      "num_tokens": 53894057.0,
+      "point_2d": 1.0,
+      "point_3d": 0.794921875,
+      "reward": 0.8730872422456741,
+      "step": 456
+    },
+    {
+      "3d_bbox": 0.7504811957478523,
+      "adv": 0.0007141208916436881,
+      "bbox": 0.9912561923265457,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.75,
+      "completions/max_terminated_length": 1594.75,
+      "completions/mean_length": 1576.1640625,
+      "completions/mean_terminated_length": 1576.1640625,
+      "completions/min_length": 1555.75,
+      "completions/min_terminated_length": 1555.75,
+      "entropy": 0.28930201614275575,
+      "epoch": 0.16372618207480594,
+      "geo": 0.8148366361856461,
+      "grad_norm": 0.275390625,
+      "learning_rate": 7.723618090452262e-06,
+      "loss": 0.0003,
+      "num_tokens": 54839099.0,
+      "point_2d": 0.875,
+      "point_3d": 0.8896484375,
+      "reward": 0.8765964508056641,
+      "step": 464
+    },
+    {
+      "3d_bbox": 0.819125022739172,
+      "adv": 0.00010633431156747974,
+      "bbox": 0.9906315729022026,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1603.0,
+      "completions/max_terminated_length": 1603.0,
+      "completions/mean_length": 1578.3203125,
+      "completions/mean_terminated_length": 1578.3203125,
+      "completions/min_length": 1556.375,
+      "completions/min_terminated_length": 1556.375,
+      "entropy": 0.28396829310804605,
+      "epoch": 0.16654904728299225,
+      "geo": 0.8597357729449868,
+      "grad_norm": 0.30078125,
+      "learning_rate": 7.683417085427137e-06,
+      "loss": 0.0004,
+      "num_tokens": 55784557.0,
+      "point_2d": 0.8916015625,
+      "point_3d": 0.9345703125,
+      "reward": 0.9089821055531502,
+      "step": 472
+    },
+    {
+      "3d_bbox": 0.7331585213541985,
+      "adv": -0.00033648866519797593,
+      "bbox": 0.8809062782675028,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.625,
+      "completions/max_terminated_length": 1594.625,
+      "completions/mean_length": 1576.11328125,
+      "completions/mean_terminated_length": 1576.11328125,
+      "completions/min_length": 1559.25,
+      "completions/min_terminated_length": 1559.25,
+      "entropy": 0.28669640561565757,
+      "epoch": 0.16937191249117856,
+      "geo": 0.9506581574678421,
+      "grad_norm": 0.2890625,
+      "learning_rate": 7.643216080402011e-06,
+      "loss": 0.0132,
+      "num_tokens": 56729514.0,
+      "point_2d": 0.8759765625,
+      "point_3d": 0.880859375,
+      "reward": 0.8427251875400543,
+      "step": 480
+    },
+    {
+      "3d_bbox": 0.7568283602595329,
+      "adv": -0.0008519564944435842,
+      "bbox": 0.9879242181777954,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1600.875,
+      "completions/max_terminated_length": 1600.875,
+      "completions/mean_length": 1574.87890625,
+      "completions/mean_terminated_length": 1574.87890625,
+      "completions/min_length": 1554.25,
+      "completions/min_terminated_length": 1554.25,
+      "entropy": 0.28737954096868634,
+      "epoch": 0.17219477769936486,
+      "geo": 0.7571075335144997,
+      "grad_norm": 0.400390625,
+      "learning_rate": 7.603015075376885e-06,
+      "loss": 0.0105,
+      "num_tokens": 57674099.0,
+      "point_2d": 0.7998046875,
+      "point_3d": 0.9228515625,
+      "reward": 0.8668522015213966,
+      "step": 488
+    },
+    {
+      "3d_bbox": 0.7658256888389587,
+      "adv": 0.0011402575073589105,
+      "bbox": 0.9932135492563248,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.25,
+      "completions/max_terminated_length": 1596.25,
+      "completions/mean_length": 1574.265625,
+      "completions/mean_terminated_length": 1574.265625,
+      "completions/min_length": 1554.375,
+      "completions/min_terminated_length": 1554.375,
+      "entropy": 0.2865150640718639,
+      "epoch": 0.17501764290755117,
+      "geo": 0.8641151636838913,
+      "grad_norm": 0.37890625,
+      "learning_rate": 7.5628140703517595e-06,
+      "loss": -0.0004,
+      "num_tokens": 58618575.0,
+      "point_2d": 0.91015625,
+      "point_3d": 0.904296875,
+      "reward": 0.8933730870485306,
+      "step": 496
+    },
+    {
+      "3d_bbox": 0.7145381905138493,
+      "adv": 0.0003104716488451231,
+      "bbox": 0.9926531538367271,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.75,
+      "completions/max_terminated_length": 1594.75,
+      "completions/mean_length": 1573.12890625,
+      "completions/mean_terminated_length": 1573.12890625,
+      "completions/min_length": 1553.375,
+      "completions/min_terminated_length": 1553.375,
+      "entropy": 0.28603550279513,
+      "epoch": 0.17784050811573748,
+      "geo": 0.7219396531581879,
+      "grad_norm": 0.310546875,
+      "learning_rate": 7.522613065326634e-06,
+      "loss": -0.0,
+      "num_tokens": 59562552.0,
+      "point_2d": 0.7509765625,
+      "point_3d": 0.9365234375,
+      "reward": 0.8486728444695473,
+      "step": 504
+    },
+    {
+      "3d_bbox": 0.7242622300982475,
+      "adv": -0.0005710294717573561,
+      "bbox": 0.9246156923472881,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1591.75,
+      "completions/max_terminated_length": 1591.75,
+      "completions/mean_length": 1571.015625,
+      "completions/mean_terminated_length": 1571.015625,
+      "completions/min_length": 1551.625,
+      "completions/min_terminated_length": 1551.625,
+      "entropy": 0.2915982282720506,
+      "epoch": 0.1806633733239238,
+      "geo": 0.7993985265493393,
+      "grad_norm": 0.408203125,
+      "learning_rate": 7.482412060301508e-06,
+      "loss": 0.0001,
+      "num_tokens": 60505948.0,
+      "point_2d": 0.7919921875,
+      "point_3d": 0.8671875,
+      "reward": 0.8270144015550613,
+      "step": 512
+    },
+    {
+      "3d_bbox": 0.7831042557954788,
+      "adv": -0.000159245344093506,
+      "bbox": 0.9960433393716812,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.375,
+      "completions/max_terminated_length": 1592.375,
+      "completions/mean_length": 1574.84375,
+      "completions/mean_terminated_length": 1574.84375,
+      "completions/min_length": 1559.125,
+      "completions/min_terminated_length": 1559.125,
+      "entropy": 0.2845937805250287,
+      "epoch": 0.1834862385321101,
+      "geo": 0.9170719608664513,
+      "grad_norm": 0.345703125,
+      "learning_rate": 7.442211055276382e-06,
+      "loss": -0.0,
+      "num_tokens": 61450564.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8408203125,
+      "reward": 0.90499197691679,
+      "step": 520
+    },
+    {
+      "3d_bbox": 0.7821980714797974,
+      "adv": -0.000531579293010509,
+      "bbox": 0.9920122027397156,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.375,
+      "completions/max_terminated_length": 1595.375,
+      "completions/mean_length": 1571.765625,
+      "completions/mean_terminated_length": 1571.765625,
+      "completions/min_length": 1551.5,
+      "completions/min_terminated_length": 1551.5,
+      "entropy": 0.2882155287079513,
+      "epoch": 0.1863091037402964,
+      "geo": 0.901853509247303,
+      "grad_norm": 0.365234375,
+      "learning_rate": 7.402010050251257e-06,
+      "loss": 0.0002,
+      "num_tokens": 62394336.0,
+      "point_2d": 0.888671875,
+      "point_3d": 0.9052734375,
+      "reward": 0.8920388966798782,
+      "step": 528
+    },
+    {
+      "3d_bbox": 0.7537402808666229,
+      "adv": -0.0004438982141437009,
+      "bbox": 0.9946581870317459,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.375,
+      "completions/max_terminated_length": 1594.375,
+      "completions/mean_length": 1574.3203125,
+      "completions/mean_terminated_length": 1574.3203125,
+      "completions/min_length": 1554.75,
+      "completions/min_terminated_length": 1554.75,
+      "entropy": 0.2886669235303998,
+      "epoch": 0.18913196894848272,
+      "geo": 0.8310712724924088,
+      "grad_norm": 0.36328125,
+      "learning_rate": 7.361809045226132e-06,
+      "loss": 0.0002,
+      "num_tokens": 63338834.0,
+      "point_2d": 0.873046875,
+      "point_3d": 0.904296875,
+      "reward": 0.8814355507493019,
+      "step": 536
+    },
+    {
+      "3d_bbox": 0.6960996612906456,
+      "adv": 0.0006180387658787367,
+      "bbox": 0.869446316966787,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1597.25,
+      "completions/max_terminated_length": 1597.25,
+      "completions/mean_length": 1571.88671875,
+      "completions/mean_terminated_length": 1571.88671875,
+      "completions/min_length": 1552.25,
+      "completions/min_terminated_length": 1552.25,
+      "entropy": 0.2887332271784544,
+      "epoch": 0.19195483415666903,
+      "geo": 0.826970137655735,
+      "grad_norm": 0.267578125,
+      "learning_rate": 7.321608040201006e-06,
+      "loss": -0.0,
+      "num_tokens": 64282685.0,
+      "point_2d": 0.875,
+      "point_3d": 0.8662109375,
+      "reward": 0.8266892246901989,
+      "step": 544
+    },
+    {
+      "3d_bbox": 0.8654715269804001,
+      "adv": -0.0008252508814621251,
+      "bbox": 0.9932455867528915,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.125,
+      "completions/max_terminated_length": 1596.125,
+      "completions/mean_length": 1572.1953125,
+      "completions/mean_terminated_length": 1572.1953125,
+      "completions/min_length": 1553.625,
+      "completions/min_terminated_length": 1553.625,
+      "entropy": 0.2873402568511665,
+      "epoch": 0.19477769936485534,
+      "geo": 0.872327484190464,
+      "grad_norm": 0.2890625,
+      "learning_rate": 7.28140703517588e-06,
+      "loss": 0.0004,
+      "num_tokens": 65226639.0,
+      "point_2d": 0.8935546875,
+      "point_3d": 0.9482421875,
+      "reward": 0.9251285046339035,
+      "step": 552
+    },
+    {
+      "3d_bbox": 0.7851101011037827,
+      "adv": -0.0011301148442726117,
+      "bbox": 0.9379025921225548,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1588.5,
+      "completions/max_terminated_length": 1588.5,
+      "completions/mean_length": 1571.4765625,
+      "completions/mean_terminated_length": 1571.4765625,
+      "completions/min_length": 1555.0,
+      "completions/min_terminated_length": 1555.0,
+      "entropy": 0.2852159277535975,
+      "epoch": 0.19760056457304165,
+      "geo": 0.690475556999445,
+      "grad_norm": 0.3125,
+      "learning_rate": 7.241206030150754e-06,
+      "loss": 0.0006,
+      "num_tokens": 66170305.0,
+      "point_2d": 0.625,
+      "point_3d": 0.837890625,
+      "reward": 0.7964758351445198,
+      "step": 560
+    },
+    {
+      "3d_bbox": 0.7313521131873131,
+      "adv": -0.000422738248744281,
+      "bbox": 0.8720479012117721,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.875,
+      "completions/max_terminated_length": 1601.875,
+      "completions/mean_length": 1574.91015625,
+      "completions/mean_terminated_length": 1574.91015625,
+      "completions/min_length": 1551.25,
+      "completions/min_terminated_length": 1551.25,
+      "entropy": 0.2880417066626251,
+      "epoch": 0.20042342978122796,
+      "geo": 0.8383861493784934,
+      "grad_norm": 0.412109375,
+      "learning_rate": 7.2010050251256295e-06,
+      "loss": 0.0004,
+      "num_tokens": 67114842.0,
+      "point_2d": 0.8740234375,
+      "point_3d": 0.84765625,
+      "reward": 0.8312699124217033,
+      "step": 568
+    },
+    {
+      "3d_bbox": 0.6841592267155647,
+      "adv": -0.0008141829575833981,
+      "bbox": 0.9068506434559822,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1597.5,
+      "completions/max_terminated_length": 1597.5,
+      "completions/mean_length": 1571.79296875,
+      "completions/mean_terminated_length": 1571.79296875,
+      "completions/min_length": 1546.75,
+      "completions/min_terminated_length": 1546.75,
+      "entropy": 0.2889366364106536,
+      "epoch": 0.20324629498941427,
+      "geo": 0.7308010272681713,
+      "grad_norm": 0.2578125,
+      "learning_rate": 7.160804020100504e-06,
+      "loss": 0.0002,
+      "num_tokens": 68058501.0,
+      "point_2d": 0.8525390625,
+      "point_3d": 0.794921875,
+      "reward": 0.8096176832914352,
+      "step": 576
+    },
+    {
+      "3d_bbox": 0.5905975513160229,
+      "adv": -0.00022596626513404772,
+      "bbox": 0.9184868931770325,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1597.5,
+      "completions/max_terminated_length": 1597.5,
+      "completions/mean_length": 1572.45703125,
+      "completions/mean_terminated_length": 1572.45703125,
+      "completions/min_length": 1551.125,
+      "completions/min_terminated_length": 1551.125,
+      "entropy": 0.2837987565435469,
+      "epoch": 0.20606916019760058,
+      "geo": 0.8267702832818031,
+      "grad_norm": 0.314453125,
+      "learning_rate": 7.120603015075378e-06,
+      "loss": -0.0003,
+      "num_tokens": 69002250.0,
+      "point_2d": 0.8681640625,
+      "point_3d": 0.7445197626948357,
+      "reward": 0.7804420664906502,
+      "step": 584
+    },
+    {
+      "3d_bbox": 0.6897582598030567,
+      "adv": -0.0004831653022847604,
+      "bbox": 0.9742325842380524,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.25,
+      "completions/max_terminated_length": 1595.25,
+      "completions/mean_length": 1572.9296875,
+      "completions/mean_terminated_length": 1572.9296875,
+      "completions/min_length": 1553.5,
+      "completions/min_terminated_length": 1553.5,
+      "entropy": 0.2835669592022896,
+      "epoch": 0.20889202540578689,
+      "geo": 0.8361655548214912,
+      "grad_norm": 0.330078125,
+      "learning_rate": 7.080402010050251e-06,
+      "loss": 0.0004,
+      "num_tokens": 69946400.0,
+      "point_2d": 0.875,
+      "point_3d": 0.8948184698820114,
+      "reward": 0.8584523350000381,
+      "step": 592
+    },
+    {
+      "3d_bbox": 0.785271767526865,
+      "adv": -0.00028875649149995297,
+      "bbox": 0.9752914533019066,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.0,
+      "completions/max_terminated_length": 1599.0,
+      "completions/mean_length": 1572.3671875,
+      "completions/mean_terminated_length": 1572.3671875,
+      "completions/min_length": 1549.875,
+      "completions/min_terminated_length": 1549.875,
+      "entropy": 0.29118923749774694,
+      "epoch": 0.2117148906139732,
+      "geo": 0.9338327869772911,
+      "grad_norm": 0.3671875,
+      "learning_rate": 7.040201005025126e-06,
+      "loss": 0.0007,
+      "num_tokens": 70890342.0,
+      "point_2d": 0.998046875,
+      "point_3d": 0.94140625,
+      "reward": 0.9250040724873543,
+      "step": 600
+    },
+    {
+      "3d_bbox": 0.8766504302620888,
+      "adv": -0.002171848598663928,
+      "bbox": 0.9872715324163437,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.125,
+      "completions/max_terminated_length": 1601.125,
+      "completions/mean_length": 1576.87890625,
+      "completions/mean_terminated_length": 1576.87890625,
+      "completions/min_length": 1556.375,
+      "completions/min_terminated_length": 1556.375,
+      "entropy": 0.28933983854949474,
+      "epoch": 0.2145377558221595,
+      "geo": 0.845995269715786,
+      "grad_norm": 0.33203125,
+      "learning_rate": 7e-06,
+      "loss": 0.0003,
+      "num_tokens": 71835359.0,
+      "point_2d": 0.875,
+      "point_3d": 0.9454848319292068,
+      "reward": 0.9211016893386841,
+      "step": 608
+    },
+    {
+      "3d_bbox": 0.8059130385518074,
+      "adv": 0.0005476666337926872,
+      "bbox": 0.9907352328300476,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1593.625,
+      "completions/max_terminated_length": 1593.625,
+      "completions/mean_length": 1570.9921875,
+      "completions/mean_terminated_length": 1570.9921875,
+      "completions/min_length": 1554.125,
+      "completions/min_terminated_length": 1554.125,
+      "entropy": 0.28401868836954236,
+      "epoch": 0.2173606210303458,
+      "geo": 0.8720198646187782,
+      "grad_norm": 0.384765625,
+      "learning_rate": 6.959798994974874e-06,
+      "loss": -0.0004,
+      "num_tokens": 72778933.0,
+      "point_2d": 0.9248046875,
+      "point_3d": 0.9135454967617989,
+      "reward": 0.9087496176362038,
+      "step": 616
+    },
+    {
+      "3d_bbox": 0.8549554497003555,
+      "adv": -0.0005420902471087174,
+      "bbox": 0.9758080616593361,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1593.0,
+      "completions/max_terminated_length": 1593.0,
+      "completions/mean_length": 1570.1640625,
+      "completions/mean_terminated_length": 1570.1640625,
+      "completions/min_length": 1549.0,
+      "completions/min_terminated_length": 1549.0,
+      "entropy": 0.29045473085716367,
+      "epoch": 0.22018348623853212,
+      "geo": 0.8916067183017731,
+      "grad_norm": 0.3828125,
+      "learning_rate": 6.919597989949749e-06,
+      "loss": 0.0156,
+      "num_tokens": 73722191.0,
+      "point_2d": 1.0,
+      "point_3d": 0.9639246314764023,
+      "reward": 0.9486720338463783,
+      "step": 624
+    },
+    {
+      "3d_bbox": 0.7975610345602036,
+      "adv": 0.0014872762099003012,
+      "bbox": 0.9193930514156818,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1600.625,
+      "completions/max_terminated_length": 1600.625,
+      "completions/mean_length": 1573.29296875,
+      "completions/mean_terminated_length": 1573.29296875,
+      "completions/min_length": 1553.0,
+      "completions/min_terminated_length": 1553.0,
+      "entropy": 0.28542913123965263,
+      "epoch": 0.22300635144671843,
+      "geo": 0.936805784702301,
+      "grad_norm": 0.2734375,
+      "learning_rate": 6.8793969849246235e-06,
+      "loss": 0.0152,
+      "num_tokens": 74666354.0,
+      "point_2d": 0.9990234375,
+      "point_3d": 0.93359375,
+      "reward": 0.9123928248882294,
+      "step": 632
+    },
+    {
+      "3d_bbox": 0.6216618679463863,
+      "adv": -0.0013801682944176719,
+      "bbox": 0.9884917959570885,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.375,
+      "completions/max_terminated_length": 1601.375,
+      "completions/mean_length": 1577.00390625,
+      "completions/mean_terminated_length": 1577.00390625,
+      "completions/min_length": 1554.0,
+      "completions/min_terminated_length": 1554.0,
+      "entropy": 0.28678226098418236,
+      "epoch": 0.22582921665490474,
+      "geo": 0.9096385985612869,
+      "grad_norm": 0.42578125,
+      "learning_rate": 6.839195979899498e-06,
+      "loss": 0.0005,
+      "num_tokens": 75611339.0,
+      "point_2d": 0.8291015625,
+      "point_3d": 0.7913028448820114,
+      "reward": 0.8076395243406296,
+      "step": 640
+    },
+    {
+      "3d_bbox": 0.7582510337233543,
+      "adv": -0.00019565537664334443,
+      "bbox": 0.9911796301603317,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.625,
+      "completions/max_terminated_length": 1596.625,
+      "completions/mean_length": 1575.71484375,
+      "completions/mean_terminated_length": 1575.71484375,
+      "completions/min_length": 1556.25,
+      "completions/min_terminated_length": 1556.25,
+      "entropy": 0.2863654876127839,
+      "epoch": 0.22865208186309105,
+      "geo": 0.9025658443570137,
+      "grad_norm": 0.451171875,
+      "learning_rate": 6.798994974874372e-06,
+      "loss": -0.0001,
+      "num_tokens": 76556042.0,
+      "point_2d": 0.95703125,
+      "point_3d": 0.8994140625,
+      "reward": 0.9014689922332764,
+      "step": 648
+    },
+    {
+      "3d_bbox": 0.753610834479332,
+      "adv": 0.00039293405916396296,
+      "bbox": 0.9021090529859066,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.625,
+      "completions/max_terminated_length": 1592.625,
+      "completions/mean_length": 1572.203125,
+      "completions/mean_terminated_length": 1572.203125,
+      "completions/min_length": 1552.25,
+      "completions/min_terminated_length": 1552.25,
+      "entropy": 0.286099414806813,
+      "epoch": 0.23147494707127736,
+      "geo": 0.9557241797447205,
+      "grad_norm": 0.455078125,
+      "learning_rate": 6.758793969849246e-06,
+      "loss": 0.0002,
+      "num_tokens": 77499886.0,
+      "point_2d": 1.0,
+      "point_3d": 0.91015625,
+      "reward": 0.8914690390229225,
+      "step": 656
+    },
+    {
+      "3d_bbox": 0.831758588552475,
+      "adv": -7.158980952226557e-05,
+      "bbox": 0.9794979766011238,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.5,
+      "completions/max_terminated_length": 1595.5,
+      "completions/mean_length": 1574.5,
+      "completions/mean_terminated_length": 1574.5,
+      "completions/min_length": 1558.75,
+      "completions/min_terminated_length": 1558.75,
+      "entropy": 0.2819104827940464,
+      "epoch": 0.23429781227946367,
+      "geo": 0.8940749987959862,
+      "grad_norm": 0.294921875,
+      "learning_rate": 6.718592964824121e-06,
+      "loss": 0.0001,
+      "num_tokens": 78444526.0,
+      "point_2d": 1.0,
+      "point_3d": 0.939453125,
+      "reward": 0.9376774355769157,
+      "step": 664
+    },
+    {
+      "3d_bbox": 0.8210256397724152,
+      "adv": 0.000540634344361024,
+      "bbox": 0.9576755836606026,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.125,
+      "completions/max_terminated_length": 1594.125,
+      "completions/mean_length": 1573.80859375,
+      "completions/mean_terminated_length": 1573.80859375,
+      "completions/min_length": 1553.25,
+      "completions/min_terminated_length": 1553.25,
+      "entropy": 0.2863646815530956,
+      "epoch": 0.23712067748764998,
+      "geo": 0.9599328562617302,
+      "grad_norm": 0.28125,
+      "learning_rate": 6.678391959798996e-06,
+      "loss": -0.0001,
+      "num_tokens": 79388949.0,
+      "point_2d": 1.0,
+      "point_3d": 0.92578125,
+      "reward": 0.9261206015944481,
+      "step": 672
+    },
+    {
+      "3d_bbox": 0.8162731677293777,
+      "adv": -0.0006006620242260396,
+      "bbox": 0.9639560282230377,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1593.75,
+      "completions/max_terminated_length": 1593.75,
+      "completions/mean_length": 1573.62890625,
+      "completions/mean_terminated_length": 1573.62890625,
+      "completions/min_length": 1557.125,
+      "completions/min_terminated_length": 1557.125,
+      "entropy": 0.2868205299600959,
+      "epoch": 0.2399435426958363,
+      "geo": 0.9127583503723145,
+      "grad_norm": 0.44921875,
+      "learning_rate": 6.63819095477387e-06,
+      "loss": -0.0,
+      "num_tokens": 80333190.0,
+      "point_2d": 1.0,
+      "point_3d": 0.9267578125,
+      "reward": 0.926746778190136,
+      "step": 680
+    },
+    {
+      "3d_bbox": 0.7580128684639931,
+      "adv": 0.0009847661577850886,
+      "bbox": 0.9946200400590897,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.0,
+      "completions/max_terminated_length": 1598.0,
+      "completions/mean_length": 1572.2109375,
+      "completions/mean_terminated_length": 1572.2109375,
+      "completions/min_length": 1552.0,
+      "completions/min_terminated_length": 1552.0,
+      "entropy": 0.286160247400403,
+      "epoch": 0.2427664079040226,
+      "geo": 0.9191197454929352,
+      "grad_norm": 0.361328125,
+      "learning_rate": 6.597989949748744e-06,
+      "loss": 0.0001,
+      "num_tokens": 81277036.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8818359375,
+      "reward": 0.9086171984672546,
+      "step": 688
+    },
+    {
+      "3d_bbox": 0.729152224957943,
+      "adv": -0.0009101523137360346,
+      "bbox": 0.9927223324775696,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.875,
+      "completions/max_terminated_length": 1592.875,
+      "completions/mean_length": 1571.8515625,
+      "completions/mean_terminated_length": 1571.8515625,
+      "completions/min_length": 1553.375,
+      "completions/min_terminated_length": 1553.375,
+      "entropy": 0.28340971609577537,
+      "epoch": 0.2455892731122089,
+      "geo": 0.9836970791220665,
+      "grad_norm": 0.34765625,
+      "learning_rate": 6.557788944723618e-06,
+      "loss": 0.0005,
+      "num_tokens": 82220822.0,
+      "point_2d": 1.0,
+      "point_3d": 0.9014246314764023,
+      "reward": 0.9058247655630112,
+      "step": 696
+    },
+    {
+      "3d_bbox": 0.7259362824261189,
+      "adv": 0.0003162266129947966,
+      "bbox": 0.922937985509634,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.0,
+      "completions/max_terminated_length": 1596.0,
+      "completions/mean_length": 1575.2265625,
+      "completions/mean_terminated_length": 1575.2265625,
+      "completions/min_length": 1555.625,
+      "completions/min_terminated_length": 1555.625,
+      "entropy": 0.286905855871737,
+      "epoch": 0.24841213832039521,
+      "geo": 0.8376269713044167,
+      "grad_norm": 0.30078125,
+      "learning_rate": 6.5175879396984935e-06,
+      "loss": 0.0,
+      "num_tokens": 83165600.0,
+      "point_2d": 0.9541015625,
+      "point_3d": 0.8759765625,
+      "reward": 0.8697381019592285,
+      "step": 704
+    },
+    {
+      "3d_bbox": 0.7864346951246262,
+      "adv": -0.0007161469466154813,
+      "bbox": 0.9958559349179268,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.0,
+      "completions/max_terminated_length": 1595.0,
+      "completions/mean_length": 1572.4921875,
+      "completions/mean_terminated_length": 1572.4921875,
+      "completions/min_length": 1545.125,
+      "completions/min_terminated_length": 1545.125,
+      "entropy": 0.2886675829067826,
+      "epoch": 0.2512350035285815,
+      "geo": 0.8416303154081106,
+      "grad_norm": 0.333984375,
+      "learning_rate": 6.477386934673368e-06,
+      "loss": 0.0008,
+      "num_tokens": 84109606.0,
+      "point_2d": 0.75,
+      "point_3d": 0.947265625,
+      "reward": 0.8698890581727028,
+      "step": 712
+    },
+    {
+      "3d_bbox": 0.657585222274065,
+      "adv": -0.0004964938416378573,
+      "bbox": 0.9940767288208008,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1589.75,
+      "completions/max_terminated_length": 1589.75,
+      "completions/mean_length": 1571.515625,
+      "completions/mean_terminated_length": 1571.515625,
+      "completions/min_length": 1553.5,
+      "completions/min_terminated_length": 1553.5,
+      "entropy": 0.2854380081407726,
+      "epoch": 0.25405786873676783,
+      "geo": 0.8985359221696854,
+      "grad_norm": 0.275390625,
+      "learning_rate": 6.437185929648242e-06,
+      "loss": 0.0,
+      "num_tokens": 85053346.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8623046875,
+      "reward": 0.8784916624426842,
+      "step": 720
+    },
+    {
+      "3d_bbox": 0.8299823924899101,
+      "adv": -0.0006786447083868552,
+      "bbox": 0.9957724660634995,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1599.75,
+      "completions/max_terminated_length": 1599.75,
+      "completions/mean_length": 1575.05078125,
+      "completions/mean_terminated_length": 1575.05078125,
+      "completions/min_length": 1553.625,
+      "completions/min_terminated_length": 1553.625,
+      "entropy": 0.2834048862569034,
+      "epoch": 0.25688073394495414,
+      "geo": 0.7720311819575727,
+      "grad_norm": 0.353515625,
+      "learning_rate": 6.396984924623116e-06,
+      "loss": 0.0008,
+      "num_tokens": 85997743.0,
+      "point_2d": 0.875,
+      "point_3d": 0.8583984375,
+      "reward": 0.8897883221507072,
+      "step": 728
+    },
+    {
+      "3d_bbox": 0.814798042178154,
+      "adv": -0.00043147145015609567,
+      "bbox": 0.994331993162632,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1603.0,
+      "completions/max_terminated_length": 1603.0,
+      "completions/mean_length": 1578.44921875,
+      "completions/mean_terminated_length": 1578.44921875,
+      "completions/min_length": 1556.25,
+      "completions/min_terminated_length": 1556.25,
+      "entropy": 0.2801376706920564,
+      "epoch": 0.25970359915314045,
+      "geo": 0.9146922156214714,
+      "grad_norm": 0.25,
+      "learning_rate": 6.3567839195979905e-06,
+      "loss": 0.0001,
+      "num_tokens": 86943082.0,
+      "point_2d": 0.8798828125,
+      "point_3d": 0.94921875,
+      "reward": 0.909557893872261,
+      "step": 736
+    },
+    {
+      "3d_bbox": 0.7457513771951199,
+      "adv": -1.4886696476423822e-05,
+      "bbox": 0.9952592924237251,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.875,
+      "completions/max_terminated_length": 1592.875,
+      "completions/mean_length": 1572.21484375,
+      "completions/mean_terminated_length": 1572.21484375,
+      "completions/min_length": 1553.625,
+      "completions/min_terminated_length": 1553.625,
+      "entropy": 0.28585455380380154,
+      "epoch": 0.26252646436132676,
+      "geo": 0.9565036073327065,
+      "grad_norm": 0.345703125,
+      "learning_rate": 6.316582914572866e-06,
+      "loss": -0.0003,
+      "num_tokens": 87886913.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8095703125,
+      "reward": 0.8876452520489693,
+      "step": 744
+    },
+    {
+      "3d_bbox": 0.6948670484125614,
+      "adv": -0.00016846299786266172,
+      "bbox": 0.8553073853254318,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.375,
+      "completions/max_terminated_length": 1594.375,
+      "completions/mean_length": 1571.93359375,
+      "completions/mean_terminated_length": 1571.93359375,
+      "completions/min_length": 1552.0,
+      "completions/min_terminated_length": 1552.0,
+      "entropy": 0.28335432009771466,
+      "epoch": 0.26534932956951307,
+      "geo": 0.7740265503525734,
+      "grad_norm": 0.421875,
+      "learning_rate": 6.27638190954774e-06,
+      "loss": 0.0162,
+      "num_tokens": 88830776.0,
+      "point_2d": 0.8701171875,
+      "point_3d": 0.9082605689764023,
+      "reward": 0.8321380317211151,
+      "step": 752
+    },
+    {
+      "3d_bbox": 0.7499241754412651,
+      "adv": 0.0011103415663455962,
+      "bbox": 0.9951865822076797,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.75,
+      "completions/max_terminated_length": 1592.75,
+      "completions/mean_length": 1570.93359375,
+      "completions/mean_terminated_length": 1570.93359375,
+      "completions/min_length": 1552.25,
+      "completions/min_terminated_length": 1552.25,
+      "entropy": 0.28731182869523764,
+      "epoch": 0.2681721947776994,
+      "geo": 0.620879240334034,
+      "grad_norm": 0.2197265625,
+      "learning_rate": 6.236180904522614e-06,
+      "loss": 0.0001,
+      "num_tokens": 89774255.0,
+      "point_2d": 0.7119140625,
+      "point_3d": 0.8916015625,
+      "reward": 0.8371566012501717,
+      "step": 760
+    },
+    {
+      "3d_bbox": 0.7572767026722431,
+      "adv": 0.0006080263119656593,
+      "bbox": 0.9954995736479759,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1589.0,
+      "completions/max_terminated_length": 1589.0,
+      "completions/mean_length": 1569.015625,
+      "completions/mean_terminated_length": 1569.015625,
+      "completions/min_length": 1549.75,
+      "completions/min_terminated_length": 1549.75,
+      "entropy": 0.2817346737720072,
+      "epoch": 0.2709950599858857,
+      "geo": 0.8461784943938255,
+      "grad_norm": 0.41015625,
+      "learning_rate": 6.195979899497488e-06,
+      "loss": -0.0,
+      "num_tokens": 90717243.0,
+      "point_2d": 1.0,
+      "point_3d": 0.892578125,
+      "reward": 0.9113386049866676,
+      "step": 768
+    },
+    {
+      "3d_bbox": 0.8124237954616547,
+      "adv": -0.00010392009426141158,
+      "bbox": 0.9695260003209114,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1590.0,
+      "completions/max_terminated_length": 1590.0,
+      "completions/mean_length": 1571.890625,
+      "completions/mean_terminated_length": 1571.890625,
+      "completions/min_length": 1553.75,
+      "completions/min_terminated_length": 1553.75,
+      "entropy": 0.2867649751715362,
+      "epoch": 0.273817925194072,
+      "geo": 0.7172370981425047,
+      "grad_norm": 0.29296875,
+      "learning_rate": 6.155778894472362e-06,
+      "loss": -0.0,
+      "num_tokens": 91660903.0,
+      "point_2d": 0.75,
+      "point_3d": 0.9462890625,
+      "reward": 0.8695597127079964,
+      "step": 776
+    },
+    {
+      "3d_bbox": 0.6247975081205368,
+      "adv": 0.00014501793702947907,
+      "bbox": 0.9793029651045799,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.125,
+      "completions/max_terminated_length": 1592.125,
+      "completions/mean_length": 1571.03515625,
+      "completions/mean_terminated_length": 1571.03515625,
+      "completions/min_length": 1550.875,
+      "completions/min_terminated_length": 1550.875,
+      "entropy": 0.2792881946079433,
+      "epoch": 0.2766407904022583,
+      "geo": 0.8185150865465403,
+      "grad_norm": 0.2890625,
+      "learning_rate": 6.115577889447236e-06,
+      "loss": 0.0,
+      "num_tokens": 92604352.0,
+      "point_2d": 0.8876953125,
+      "point_3d": 0.8994140625,
+      "reward": 0.8478024452924728,
+      "step": 784
+    },
+    {
+      "3d_bbox": 0.7522557079792023,
+      "adv": 0.00011037173771910602,
+      "bbox": 0.8692253082990646,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.625,
+      "completions/max_terminated_length": 1594.625,
+      "completions/mean_length": 1572.27734375,
+      "completions/mean_terminated_length": 1572.27734375,
+      "completions/min_length": 1552.5,
+      "completions/min_terminated_length": 1552.5,
+      "entropy": 0.2874673060141504,
+      "epoch": 0.2794636556104446,
+      "geo": 0.9408122524619102,
+      "grad_norm": 0.31640625,
+      "learning_rate": 6.07537688442211e-06,
+      "loss": 0.0144,
+      "num_tokens": 93548159.0,
+      "point_2d": 0.7587890625,
+      "point_3d": 0.91796875,
+      "reward": 0.8245597034692764,
+      "step": 792
+    },
+    {
+      "3d_bbox": 0.8205922544002533,
+      "adv": 0.0001977139909286052,
+      "bbox": 0.9971112534403801,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1589.25,
+      "completions/max_terminated_length": 1589.25,
+      "completions/mean_length": 1571.43359375,
+      "completions/mean_terminated_length": 1571.43359375,
+      "completions/min_length": 1552.625,
+      "completions/min_terminated_length": 1552.625,
+      "entropy": 0.28211602522060275,
+      "epoch": 0.2822865208186309,
+      "geo": 0.8445243984460831,
+      "grad_norm": 0.3203125,
+      "learning_rate": 6.035175879396985e-06,
+      "loss": 0.0007,
+      "num_tokens": 94491662.0,
+      "point_2d": 0.884765625,
+      "point_3d": 0.9365234375,
+      "reward": 0.9097481444478035,
+      "step": 800
+    },
+    {
+      "3d_bbox": 0.794718824326992,
+      "adv": -0.00023952151968842372,
+      "bbox": 0.9944768622517586,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1601.375,
+      "completions/max_terminated_length": 1601.375,
+      "completions/mean_length": 1571.67578125,
+      "completions/mean_terminated_length": 1571.67578125,
+      "completions/min_length": 1552.125,
+      "completions/min_terminated_length": 1552.125,
+      "entropy": 0.2849376816302538,
+      "epoch": 0.28510938602681724,
+      "geo": 0.9431384205818176,
+      "grad_norm": 0.34375,
+      "learning_rate": 5.99497487437186e-06,
+      "loss": 0.0001,
+      "num_tokens": 95435451.0,
+      "point_2d": 0.9990234375,
+      "point_3d": 0.8720703125,
+      "reward": 0.9150723442435265,
+      "step": 808
+    },
+    {
+      "3d_bbox": 0.7696447446942329,
+      "adv": 0.0008425721316598356,
+      "bbox": 0.9485463574528694,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1593.875,
+      "completions/max_terminated_length": 1593.875,
+      "completions/mean_length": 1570.11328125,
+      "completions/mean_terminated_length": 1570.11328125,
+      "completions/min_length": 1553.875,
+      "completions/min_terminated_length": 1553.875,
+      "entropy": 0.28682070411741734,
+      "epoch": 0.28793225123500354,
+      "geo": 0.9097637981176376,
+      "grad_norm": 0.2353515625,
+      "learning_rate": 5.954773869346734e-06,
+      "loss": -0.0006,
+      "num_tokens": 96378680.0,
+      "point_2d": 0.9990234375,
+      "point_3d": 0.9013671875,
+      "reward": 0.9046454280614853,
+      "step": 816
+    },
+    {
+      "3d_bbox": 0.6897615268826485,
+      "adv": -0.000497654465107189,
+      "bbox": 0.8617971315979958,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1602.125,
+      "completions/max_terminated_length": 1602.125,
+      "completions/mean_length": 1577.7734375,
+      "completions/mean_terminated_length": 1577.7734375,
+      "completions/min_length": 1558.125,
+      "completions/min_terminated_length": 1558.125,
+      "entropy": 0.28228796226903796,
+      "epoch": 0.29075511644318985,
+      "geo": 0.7427511867135763,
+      "grad_norm": 0.326171875,
+      "learning_rate": 5.914572864321608e-06,
+      "loss": 0.0001,
+      "num_tokens": 97324310.0,
+      "point_2d": 0.7529296875,
+      "point_3d": 0.919921875,
+      "reward": 0.8061025515198708,
+      "step": 824
+    },
+    {
+      "3d_bbox": 0.6537492573261261,
+      "adv": -0.00023956477116371389,
+      "bbox": 0.9781710132956505,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.0,
+      "completions/max_terminated_length": 1596.0,
+      "completions/mean_length": 1574.296875,
+      "completions/mean_terminated_length": 1574.296875,
+      "completions/min_length": 1554.75,
+      "completions/min_terminated_length": 1554.75,
+      "entropy": 0.28043507505208254,
+      "epoch": 0.29357798165137616,
+      "geo": 0.8402111828327179,
+      "grad_norm": 0.291015625,
+      "learning_rate": 5.874371859296482e-06,
+      "loss": -0.0001,
+      "num_tokens": 98268682.0,
+      "point_2d": 1.0,
+      "point_3d": 0.9551355689764023,
+      "reward": 0.8967639729380608,
+      "step": 832
+    },
+    {
+      "3d_bbox": 0.755732849240303,
+      "adv": -0.000661049871268915,
+      "bbox": 0.9913836792111397,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1598.125,
+      "completions/max_terminated_length": 1598.125,
+      "completions/mean_length": 1575.3125,
+      "completions/mean_terminated_length": 1575.3125,
+      "completions/min_length": 1555.625,
+      "completions/min_terminated_length": 1555.625,
+      "entropy": 0.2857117219828069,
+      "epoch": 0.2964008468595625,
+      "geo": 0.8524684980511665,
+      "grad_norm": 0.380859375,
+      "learning_rate": 5.8341708542713575e-06,
+      "loss": 0.0143,
+      "num_tokens": 99213386.0,
+      "point_2d": 0.875,
+      "point_3d": 0.876953125,
+      "reward": 0.8747674003243446,
+      "step": 840
+    },
+    {
+      "3d_bbox": 0.7413047328591347,
+      "adv": 0.0002993446832988411,
+      "bbox": 0.8264872687868774,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.125,
+      "completions/max_terminated_length": 1594.125,
+      "completions/mean_length": 1573.984375,
+      "completions/mean_terminated_length": 1573.984375,
+      "completions/min_length": 1556.75,
+      "completions/min_terminated_length": 1556.75,
+      "entropy": 0.28746917750686407,
+      "epoch": 0.2992237120677488,
+      "geo": 0.9443965330719948,
+      "grad_norm": 0.3359375,
+      "learning_rate": 5.793969849246232e-06,
+      "loss": 0.0195,
+      "num_tokens": 100157950.0,
+      "point_2d": 1.0,
+      "point_3d": 0.9140625,
+      "reward": 0.8704636245965958,
+      "step": 848
+    },
+    {
+      "3d_bbox": 0.7137239277362823,
+      "adv": -0.06107998272636905,
+      "bbox": 0.9737529084086418,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1593.25,
+      "completions/max_terminated_length": 1593.25,
+      "completions/mean_length": 1571.8046875,
+      "completions/mean_terminated_length": 1571.8046875,
+      "completions/min_length": 1553.5,
+      "completions/min_terminated_length": 1553.5,
+      "entropy": 0.2830418418161571,
+      "epoch": 0.3020465772759351,
+      "geo": 0.7549894053954631,
+      "grad_norm": 0.3671875,
+      "learning_rate": 5.753768844221106e-06,
+      "loss": 0.0151,
+      "num_tokens": 101101604.0,
+      "point_2d": 0.7958984375,
+      "point_3d": 0.865234375,
+      "reward": 0.837152399122715,
+      "step": 856
+    },
+    {
+      "3d_bbox": 0.6744807064533234,
+      "adv": 0.0013020644328207709,
+      "bbox": 0.9808441251516342,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.625,
+      "completions/max_terminated_length": 1596.625,
+      "completions/mean_length": 1574.37109375,
+      "completions/mean_terminated_length": 1574.37109375,
+      "completions/min_length": 1552.875,
+      "completions/min_terminated_length": 1552.875,
+      "entropy": 0.28136514313519,
+      "epoch": 0.3048694424841214,
+      "geo": 0.7662752494215965,
+      "grad_norm": 0.333984375,
+      "learning_rate": 5.71356783919598e-06,
+      "loss": -0.0,
+      "num_tokens": 102045955.0,
+      "point_2d": 0.77734375,
+      "point_3d": 0.91015625,
+      "reward": 0.8357062116265297,
+      "step": 864
+    },
+    {
+      "3d_bbox": 0.7125367447733879,
+      "adv": 0.0008483012829856307,
+      "bbox": 0.8669970044866204,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1593.375,
+      "completions/max_terminated_length": 1593.375,
+      "completions/mean_length": 1573.75390625,
+      "completions/mean_terminated_length": 1573.75390625,
+      "completions/min_length": 1556.5,
+      "completions/min_terminated_length": 1556.5,
+      "entropy": 0.2820960213430226,
+      "epoch": 0.3076923076923077,
+      "geo": 0.6582350172102451,
+      "grad_norm": 0.296875,
+      "learning_rate": 5.6733668341708545e-06,
+      "loss": 0.0001,
+      "num_tokens": 102990100.0,
+      "point_2d": 0.5009765625,
+      "point_3d": 0.896484375,
+      "reward": 0.7442486621439457,
+      "step": 872
+    },
+    {
+      "3d_bbox": 0.8094568699598312,
+      "adv": 0.000244778927481093,
+      "bbox": 0.8707572477287613,
+      "completions/clipped_ratio": 0.00390625,
+      "completions/max_length": 1654.875,
+      "completions/max_terminated_length": 1598.75,
+      "completions/mean_length": 1576.1015625,
+      "completions/mean_terminated_length": 1574.2484893798828,
+      "completions/min_length": 1555.0,
+      "completions/min_terminated_length": 1555.0,
+      "entropy": 0.28441567649133503,
+      "epoch": 0.310515172900494,
+      "geo": 0.9321269914507866,
+      "grad_norm": 0.5546875,
+      "learning_rate": 5.633165829145729e-06,
+      "loss": 0.026,
+      "num_tokens": 103935062.0,
+      "point_2d": 0.8828125,
+      "point_3d": 0.9479817748069763,
+      "reward": 0.8777521066367626,
+      "step": 880
+    },
+    {
+      "3d_bbox": 0.6922821514308453,
+      "adv": -7.48095062590437e-05,
+      "bbox": 0.9605405703186989,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.25,
+      "completions/max_terminated_length": 1596.25,
+      "completions/mean_length": 1574.234375,
+      "completions/mean_terminated_length": 1574.234375,
+      "completions/min_length": 1555.875,
+      "completions/min_terminated_length": 1555.875,
+      "entropy": 0.28310656640678644,
+      "epoch": 0.31333803810868033,
+      "geo": 0.9695179983973503,
+      "grad_norm": 0.283203125,
+      "learning_rate": 5.592964824120604e-06,
+      "loss": 0.0,
+      "num_tokens": 104879434.0,
+      "point_2d": 0.9921875,
+      "point_3d": 0.921875,
+      "reward": 0.8917213082313538,
+      "step": 888
+    },
+    {
+      "3d_bbox": 0.766367569565773,
+      "adv": -0.0006123652386804679,
+      "bbox": 0.9947258159518242,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.0,
+      "completions/max_terminated_length": 1594.0,
+      "completions/mean_length": 1572.81640625,
+      "completions/mean_terminated_length": 1572.81640625,
+      "completions/min_length": 1554.125,
+      "completions/min_terminated_length": 1554.125,
+      "entropy": 0.2817820068448782,
+      "epoch": 0.31616090331686664,
+      "geo": 0.9325108900666237,
+      "grad_norm": 0.359375,
+      "learning_rate": 5.552763819095478e-06,
+      "loss": 0.0001,
+      "num_tokens": 105823315.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8896484375,
+      "reward": 0.9126854613423347,
+      "step": 896
+    },
+    {
+      "3d_bbox": 0.7490993067622185,
+      "adv": -0.00017882781867228914,
+      "bbox": 0.9934130385518074,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1592.75,
+      "completions/max_terminated_length": 1592.75,
+      "completions/mean_length": 1572.3125,
+      "completions/mean_terminated_length": 1572.3125,
+      "completions/min_length": 1550.875,
+      "completions/min_terminated_length": 1550.875,
+      "entropy": 0.2811024447437376,
+      "epoch": 0.31898376852505295,
+      "geo": 0.8058975423919037,
+      "grad_norm": 0.333984375,
+      "learning_rate": 5.512562814070352e-06,
+      "loss": 0.0001,
+      "num_tokens": 106767163.0,
+      "point_2d": 0.875,
+      "point_3d": 0.880859375,
+      "reward": 0.8745929300785065,
+      "step": 904
+    },
+    {
+      "3d_bbox": 0.690778411924839,
+      "adv": -0.0004784437787748175,
+      "bbox": 0.9014169201254845,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1600.875,
+      "completions/max_terminated_length": 1600.875,
+      "completions/mean_length": 1574.25,
+      "completions/mean_terminated_length": 1574.25,
+      "completions/min_length": 1551.125,
+      "completions/min_terminated_length": 1551.125,
+      "entropy": 0.2825395744293928,
+      "epoch": 0.32180663373323926,
+      "geo": 0.654383197426796,
+      "grad_norm": 0.373046875,
+      "learning_rate": 5.472361809045227e-06,
+      "loss": 0.0005,
+      "num_tokens": 107711403.0,
+      "point_2d": 0.7802734375,
+      "point_3d": 0.9189453125,
+      "reward": 0.8228535130620003,
+      "step": 912
+    },
+    {
+      "3d_bbox": 0.8111849427223206,
+      "adv": -0.0003063662707063486,
+      "bbox": 0.960642546415329,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1596.125,
+      "completions/max_terminated_length": 1596.125,
+      "completions/mean_length": 1575.28125,
+      "completions/mean_terminated_length": 1575.28125,
+      "completions/min_length": 1555.75,
+      "completions/min_terminated_length": 1555.75,
+      "entropy": 0.28817785205319524,
+      "epoch": 0.32462949894142556,
+      "geo": 0.7242828160524368,
+      "grad_norm": 0.40234375,
+      "learning_rate": 5.432160804020101e-06,
+      "loss": 0.0148,
+      "num_tokens": 108655899.0,
+      "point_2d": 0.75,
+      "point_3d": 0.9453125,
+      "reward": 0.866785004734993,
+      "step": 920
+    },
+    {
+      "3d_bbox": 0.80051339417696,
+      "adv": -0.0008653635522932746,
+      "bbox": 0.9517601355910301,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1604.125,
+      "completions/max_terminated_length": 1604.125,
+      "completions/mean_length": 1580.0546875,
+      "completions/mean_terminated_length": 1580.0546875,
+      "completions/min_length": 1561.25,
+      "completions/min_terminated_length": 1561.25,
+      "entropy": 0.2851322675123811,
+      "epoch": 0.3274523641496119,
+      "geo": 0.9541479349136353,
+      "grad_norm": 0.404296875,
+      "learning_rate": 5.391959798994976e-06,
+      "loss": 0.0006,
+      "num_tokens": 109602009.0,
+      "point_2d": 1.0,
+      "point_3d": 0.9443359375,
+      "reward": 0.9241523742675781,
+      "step": 928
+    },
+    {
+      "3d_bbox": 0.7602918669581413,
+      "adv": 0.0012147135075792903,
+      "bbox": 0.9494145661592484,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1595.875,
+      "completions/max_terminated_length": 1595.875,
+      "completions/mean_length": 1576.296875,
+      "completions/mean_terminated_length": 1576.296875,
+      "completions/min_length": 1557.125,
+      "completions/min_terminated_length": 1557.125,
+      "entropy": 0.2826410192064941,
+      "epoch": 0.3302752293577982,
+      "geo": 0.8968632370233536,
+      "grad_norm": 0.2734375,
+      "learning_rate": 5.35175879396985e-06,
+      "loss": -0.0001,
+      "num_tokens": 110546973.0,
+      "point_2d": 1.0,
+      "point_3d": 0.9013671875,
+      "reward": 0.9027684107422829,
+      "step": 936
+    },
+    {
+      "3d_bbox": 0.6780652180314064,
+      "adv": -0.0008688528978382237,
+      "bbox": 0.9930703118443489,
+      "completions/clipped_ratio": 0.0,
+      "completions/max_length": 1594.625,
+      "completions/max_terminated_length": 1594.625,
+      "completions/mean_length": 1574.8828125,
+      "completions/mean_terminated_length": 1574.8828125,
+      "completions/min_length": 1557.125,
+      "completions/min_terminated_length": 1557.125,
+      "entropy": 0.2831588592380285,
+      "epoch": 0.3330980945659845,
+      "geo": 0.9448152706027031,
+      "grad_norm": 0.26953125,
+      "learning_rate": 5.3115577889447245e-06,
+      "loss": -0.0,
+      "num_tokens": 111491407.0,
+      "point_2d": 1.0,
+      "point_3d": 0.8857421875,
+      "reward": 0.8892194256186485,
+      "step": 944
+    }
+  ],
+  "logging_steps": 8,
+  "max_steps": 2000,
+  "num_input_tokens_seen": 112199787,
+  "num_train_epochs": 1,
+  "save_steps": 50,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-950/video_preprocessor_config.json

@@ -0,0 +1,44 @@
+{
+  "crop_size": null,
+  "data_format": "channels_first",
+  "default_to_square": true,
+  "device": null,
+  "do_center_crop": null,
+  "do_convert_rgb": true,
+  "do_normalize": true,
+  "do_pad": null,
+  "do_rescale": true,
+  "do_resize": true,
+  "do_sample_frames": false,
+  "fps": null,
+  "image_mean": [
+    0.48145466,
+    0.4578275,
+    0.40821073
+  ],
+  "image_std": [
+    0.26862954,
+    0.26130258,
+    0.27577711
+  ],
+  "input_data_format": null,
+  "max_frames": 768,
+  "max_pixels": 12845056,
+  "merge_size": 2,
+  "min_frames": 4,
+  "min_pixels": 3136,
+  "num_frames": null,
+  "patch_size": 14,
+  "processor_class": "Qwen2_5_VLProcessor",
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "return_metadata": false,
+  "size": {
+    "longest_edge": 12845056,
+    "shortest_edge": 3136
+  },
+  "size_divisor": null,
+  "temporal_patch_size": 2,
+  "video_metadata": null,
+  "video_processor_type": "Qwen2VLVideoProcessor"
+}