metadata
dataset_info:
- config_name: gemma-4-26b-a4b-it-ins-gen-2048-vt280
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 91523033
num_examples: 1000
download_size: 90535295
dataset_size: 91523033
- config_name: gemma-4-26b-a4b-it-ins-gen-2048-vt280-subsample-50
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 5928740
num_examples: 50
download_size: 5055620
dataset_size: 5928740
- config_name: gemma-4-26b-a4b-it-ins-gen-2048-vt280-subsample-100
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 9582264
num_examples: 100
download_size: 7601049
dataset_size: 9582264
- config_name: gemma-4-26b-a4b-it-think-gen-4096-vt280
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 101044267
num_examples: 993
download_size: 99518233
dataset_size: 101044267
- config_name: gemma-4-26b-a4b-it-think-gen-4096-vt280-subsample-50
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 6185435
num_examples: 50
download_size: 4484151
dataset_size: 6185435
- config_name: gemma-4-26b-a4b-it-think-gen-4096-vt280-subsample-100
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 11142397
num_examples: 100
download_size: 8482300
dataset_size: 11142397
- config_name: gemma-4-e4b-it-ins-gen-2048-vt140
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 89622333
num_examples: 1000
download_size: 88608415
dataset_size: 89622333
- config_name: gemma-4-e4b-it-ins-gen-2048-vt140-subsample-50
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 5825325
num_examples: 50
download_size: 5056883
dataset_size: 5825325
- config_name: gemma-4-e4b-it-ins-gen-2048-vt140-subsample-100
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 9386809
num_examples: 100
download_size: 7607072
dataset_size: 9386809
- config_name: gemma-4-e4b-it-ins-gen-2048-vt280
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 91756999
num_examples: 1000
download_size: 90746783
dataset_size: 91756999
- config_name: gemma-4-e4b-it-ins-gen-2048-vt280-subsample-50
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 5948978
num_examples: 50
download_size: 5061050
dataset_size: 5948978
- config_name: gemma-4-e4b-it-ins-gen-2048-vt280-subsample-100
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 9618072
num_examples: 100
download_size: 7611493
dataset_size: 9618072
- config_name: gemma-4-e4b-it-ins-gen-2048-vt560
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 96112223
num_examples: 1000
download_size: 95105165
dataset_size: 96112223
- config_name: gemma-4-e4b-it-ins-gen-2048-vt560-subsample-50
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 6165346
num_examples: 50
download_size: 5059196
dataset_size: 6165346
- config_name: gemma-4-e4b-it-ins-gen-2048-vt560-subsample-100
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 10063890
num_examples: 100
download_size: 7613662
dataset_size: 10063890
- config_name: gemma-4-e4b-it-think-gen-4096-vt140
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 93515185
num_examples: 999
download_size: 92022189
dataset_size: 93515185
- config_name: gemma-4-e4b-it-think-gen-4096-vt140-subsample-50
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 3869496
num_examples: 50
download_size: 3256474
dataset_size: 3869496
- config_name: gemma-4-e4b-it-think-gen-4096-vt140-subsample-100
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 8722186
num_examples: 100
download_size: 6848133
dataset_size: 8722186
- config_name: gemma-4-e4b-it-think-gen-4096-vt280
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 95960622
num_examples: 1000
download_size: 94481110
dataset_size: 95960622
- config_name: gemma-4-e4b-it-think-gen-4096-vt280-subsample-50
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 6226932
num_examples: 50
download_size: 5110915
dataset_size: 6226932
- config_name: gemma-4-e4b-it-think-gen-4096-vt280-subsample-100
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 10150356
num_examples: 100
download_size: 7702083
dataset_size: 10150356
- config_name: gemma-4-e4b-it-think-gen-4096-vt560
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 100109306
num_examples: 1000
download_size: 98680459
dataset_size: 100109306
- config_name: gemma-4-e4b-it-think-gen-4096-vt560-subsample-50
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 6455712
num_examples: 50
download_size: 5119850
dataset_size: 6455712
- config_name: gemma-4-e4b-it-think-gen-4096-vt560-subsample-100
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 10592198
num_examples: 100
download_size: 7703716
dataset_size: 10592198
- config_name: qwen3.5-4b-ins-gen-2048
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 100850370
num_examples: 1000
download_size: 100049114
dataset_size: 100850370
- config_name: qwen3.5-4b-ins-gen-2048-subsample-50
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 6512026
num_examples: 50
download_size: 5065405
dataset_size: 6512026
- config_name: qwen3.5-4b-ins-gen-2048-subsample-100
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 10590172
num_examples: 100
download_size: 7621616
dataset_size: 10590172
- config_name: qwen3.5-4b-think-gen-4096
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 116627550
num_examples: 996
download_size: 114686446
dataset_size: 116627550
- config_name: qwen3.5-4b-think-gen-4096-subsample-50
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 9470444
num_examples: 50
download_size: 6227347
dataset_size: 9470444
- config_name: qwen3.5-4b-think-gen-4096-subsample-100
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 14164511
num_examples: 100
download_size: 8996353
dataset_size: 14164511
- config_name: qwen3.6-35b-a3b-ins-gen-2048
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 100801221
num_examples: 1000
download_size: 99999534
dataset_size: 100801221
- config_name: qwen3.6-35b-a3b-ins-gen-2048-subsample-50
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 6519555
num_examples: 50
download_size: 5066949
dataset_size: 6519555
- config_name: qwen3.6-35b-a3b-ins-gen-2048-subsample-100
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 10584317
num_examples: 100
download_size: 7620096
dataset_size: 10584317
- config_name: qwen3.6-35b-a3b-think-gen-4096
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 112098067
num_examples: 999
download_size: 110511994
dataset_size: 112098067
- config_name: qwen3.6-35b-a3b-think-gen-4096-subsample-50
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 4862967
num_examples: 50
download_size: 3528184
dataset_size: 4862967
- config_name: qwen3.6-35b-a3b-think-gen-4096-subsample-100
features:
- name: source
dtype: string
- name: item_id
dtype: string
- name: origin_id
dtype: int64
- name: sample_seed
dtype: int64
- name: question
dtype: string
- name: images
list: image
- name: num_images
dtype: int32
- name: ref_answer
dtype: string
- name: category
dtype: string
- name: input_ids
list: int64
- name: input_tokens_len
dtype: int64
- name: generated_texts
dtype: string
- name: generated_tokens_len
dtype: int64
- name: n_prefill_tokens
dtype: int64
- name: finish_reason
dtype: string
- name: seed
dtype: int64
- name: labels
list: int64
- name: image_fingerprint
dtype: string
- name: image_processor_config
dtype: string
- name: image_processor_config_hash
dtype: string
- name: sum_vision_tokens
dtype: int64
- name: max_vision_tokens
dtype: int64
- name: id
dtype: string
- name: truncated_by_cap
dtype: bool
- name: generation_model_name_or_path
dtype: string
- name: enable_thinking
dtype: bool
splits:
- name: train
num_bytes: 11689761
num_examples: 100
download_size: 7894692
dataset_size: 11689761
configs:
- config_name: gemma-4-26b-a4b-it-ins-gen-2048-vt280
data_files:
- split: train
path: gemma-4-26b-a4b-it-ins-gen-2048-vt280/train-*
- config_name: gemma-4-26b-a4b-it-ins-gen-2048-vt280-subsample-50
data_files:
- split: train
path: gemma-4-26b-a4b-it-ins-gen-2048-vt280-subsample-50/train-*
- config_name: gemma-4-26b-a4b-it-ins-gen-2048-vt280-subsample-100
data_files:
- split: train
path: gemma-4-26b-a4b-it-ins-gen-2048-vt280-subsample-100/train-*
- config_name: gemma-4-26b-a4b-it-think-gen-4096-vt280
data_files:
- split: train
path: gemma-4-26b-a4b-it-think-gen-4096-vt280/train-*
- config_name: gemma-4-26b-a4b-it-think-gen-4096-vt280-subsample-50
data_files:
- split: train
path: gemma-4-26b-a4b-it-think-gen-4096-vt280-subsample-50/train-*
- config_name: gemma-4-26b-a4b-it-think-gen-4096-vt280-subsample-100
data_files:
- split: train
path: gemma-4-26b-a4b-it-think-gen-4096-vt280-subsample-100/train-*
- config_name: gemma-4-e4b-it-ins-gen-2048-vt140
data_files:
- split: train
path: gemma-4-e4b-it-ins-gen-2048-vt140/train-*
- config_name: gemma-4-e4b-it-ins-gen-2048-vt140-subsample-50
data_files:
- split: train
path: gemma-4-e4b-it-ins-gen-2048-vt140-subsample-50/train-*
- config_name: gemma-4-e4b-it-ins-gen-2048-vt140-subsample-100
data_files:
- split: train
path: gemma-4-e4b-it-ins-gen-2048-vt140-subsample-100/train-*
- config_name: gemma-4-e4b-it-ins-gen-2048-vt280
data_files:
- split: train
path: gemma-4-e4b-it-ins-gen-2048-vt280/train-*
- config_name: gemma-4-e4b-it-ins-gen-2048-vt280-subsample-50
data_files:
- split: train
path: gemma-4-e4b-it-ins-gen-2048-vt280-subsample-50/train-*
- config_name: gemma-4-e4b-it-ins-gen-2048-vt280-subsample-100
data_files:
- split: train
path: gemma-4-e4b-it-ins-gen-2048-vt280-subsample-100/train-*
- config_name: gemma-4-e4b-it-ins-gen-2048-vt560
data_files:
- split: train
path: gemma-4-e4b-it-ins-gen-2048-vt560/train-*
- config_name: gemma-4-e4b-it-ins-gen-2048-vt560-subsample-50
data_files:
- split: train
path: gemma-4-e4b-it-ins-gen-2048-vt560-subsample-50/train-*
- config_name: gemma-4-e4b-it-ins-gen-2048-vt560-subsample-100
data_files:
- split: train
path: gemma-4-e4b-it-ins-gen-2048-vt560-subsample-100/train-*
- config_name: gemma-4-e4b-it-think-gen-4096-vt140
data_files:
- split: train
path: gemma-4-e4b-it-think-gen-4096-vt140/train-*
- config_name: gemma-4-e4b-it-think-gen-4096-vt140-subsample-50
data_files:
- split: train
path: gemma-4-e4b-it-think-gen-4096-vt140-subsample-50/train-*
- config_name: gemma-4-e4b-it-think-gen-4096-vt140-subsample-100
data_files:
- split: train
path: gemma-4-e4b-it-think-gen-4096-vt140-subsample-100/train-*
- config_name: gemma-4-e4b-it-think-gen-4096-vt280
data_files:
- split: train
path: gemma-4-e4b-it-think-gen-4096-vt280/train-*
- config_name: gemma-4-e4b-it-think-gen-4096-vt280-subsample-50
data_files:
- split: train
path: gemma-4-e4b-it-think-gen-4096-vt280-subsample-50/train-*
- config_name: gemma-4-e4b-it-think-gen-4096-vt280-subsample-100
data_files:
- split: train
path: gemma-4-e4b-it-think-gen-4096-vt280-subsample-100/train-*
- config_name: gemma-4-e4b-it-think-gen-4096-vt560
data_files:
- split: train
path: gemma-4-e4b-it-think-gen-4096-vt560/train-*
- config_name: gemma-4-e4b-it-think-gen-4096-vt560-subsample-50
data_files:
- split: train
path: gemma-4-e4b-it-think-gen-4096-vt560-subsample-50/train-*
- config_name: gemma-4-e4b-it-think-gen-4096-vt560-subsample-100
data_files:
- split: train
path: gemma-4-e4b-it-think-gen-4096-vt560-subsample-100/train-*
- config_name: qwen3.5-4b-ins-gen-2048
data_files:
- split: train
path: qwen3.5-4b-ins-gen-2048/train-*
- config_name: qwen3.5-4b-ins-gen-2048-subsample-50
data_files:
- split: train
path: qwen3.5-4b-ins-gen-2048-subsample-50/train-*
- config_name: qwen3.5-4b-ins-gen-2048-subsample-100
data_files:
- split: train
path: qwen3.5-4b-ins-gen-2048-subsample-100/train-*
- config_name: qwen3.5-4b-think-gen-4096
data_files:
- split: train
path: qwen3.5-4b-think-gen-4096/train-*
- config_name: qwen3.5-4b-think-gen-4096-subsample-50
data_files:
- split: train
path: qwen3.5-4b-think-gen-4096-subsample-50/train-*
- config_name: qwen3.5-4b-think-gen-4096-subsample-100
data_files:
- split: train
path: qwen3.5-4b-think-gen-4096-subsample-100/train-*
- config_name: qwen3.6-35b-a3b-ins-gen-2048
data_files:
- split: train
path: qwen3.6-35b-a3b-ins-gen-2048/train-*
- config_name: qwen3.6-35b-a3b-ins-gen-2048-subsample-50
data_files:
- split: train
path: qwen3.6-35b-a3b-ins-gen-2048-subsample-50/train-*
- config_name: qwen3.6-35b-a3b-ins-gen-2048-subsample-100
data_files:
- split: train
path: qwen3.6-35b-a3b-ins-gen-2048-subsample-100/train-*
- config_name: qwen3.6-35b-a3b-think-gen-4096
data_files:
- split: train
path: qwen3.6-35b-a3b-think-gen-4096/train-*
- config_name: qwen3.6-35b-a3b-think-gen-4096-subsample-50
data_files:
- split: train
path: qwen3.6-35b-a3b-think-gen-4096-subsample-50/train-*
- config_name: qwen3.6-35b-a3b-think-gen-4096-subsample-100
data_files:
- split: train
path: qwen3.6-35b-a3b-think-gen-4096-subsample-100/train-*