Datasets:

True2456
/

iq-terrain-vlm-dataset

Error code:   DatasetGenerationError
Exception:    TypeError
Message:      Couldn't cast array of type
struct<role: string, content: string, tools: list<item: struct<type: string, function: struct<name: string, description: string, parameters: struct<type: string, properties: struct<technique: struct<type: string, enum: list<item: string>>, octaves: struct<type: string, minimum: int64, maximum: int64>, lacunarity: struct<type: string>, gain: struct<type: string>, resolution: struct<type: string, items: struct<type: string>>, camera_mode: struct<type: string, enum: list<item: string>>, palette: struct<type: string>, seed: struct<type: string>, effect: struct<type: string, enum: list<item: string>>, penumbra_k: struct<type: string>, fog_density: struct<type: string>, height_decay: struct<type: string>, light_dir: struct<type: string, items: struct<type: string>>, ambient_intensity: struct<type: string>, shadow_steps: struct<type: string>, operation: struct<type: string, enum: list<item: string>>, blend_factor_k: struct<type: string>, primitive_a: struct<type: string>, primitive_b: struct<type: string>, displacement_scale: struct<type: string>, shader_type: struct<type: string, enum: list<item: string>>, target: struct<type: string, enum: list<item: string>>, includes: struct<type: string, items: struct<type: string>>, output_format: struct<type: string, enum: list<item: string>>>, required: list<item: string>, additionalProperties: bool>>>>, tool_calls: list<item: struct<id: string, type: string, function: struct<name: string, arguments: string>>>>
to
{'role': Value('string'), 'content': Value('string')}
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1816, in _prepare_split_single
                  for key, table in generator:
                                    ^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 613, in wrapped
                  for item in generator(*args, **kwargs):
                              ~~~~~~~~~^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2369, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2303, in cast_table_to_schema
                  cast_array_to_feature(
                  ~~~~~~~~~~~~~~~~~~~~~^
                      table[name] if name in table_column_names else pa.array([None] * len(table), type=schema.field(name).type),
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                      feature,
                      ^^^^^^^^
                  )
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 1852, in wrapper
                  return pa.chunked_array([func(chunk, *args, **kwargs) for chunk in array.chunks])
                                           ~~~~^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2109, in cast_array_to_feature
                  casted_array_values = _c(array.values, feature.feature)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 1854, in wrapper
                  return func(array, *args, **kwargs)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2149, in cast_array_to_feature
                  raise TypeError(f"Couldn't cast array of type\n{_short_str(array.type)}\nto\n{_short_str(feature)}")
              TypeError: Couldn't cast array of type
              struct<role: string, content: string, tools: list<item: struct<type: string, function: struct<name: string, description: string, parameters: struct<type: string, properties: struct<technique: struct<type: string, enum: list<item: string>>, octaves: struct<type: string, minimum: int64, maximum: int64>, lacunarity: struct<type: string>, gain: struct<type: string>, resolution: struct<type: string, items: struct<type: string>>, camera_mode: struct<type: string, enum: list<item: string>>, palette: struct<type: string>, seed: struct<type: string>, effect: struct<type: string, enum: list<item: string>>, penumbra_k: struct<type: string>, fog_density: struct<type: string>, height_decay: struct<type: string>, light_dir: struct<type: string, items: struct<type: string>>, ambient_intensity: struct<type: string>, shadow_steps: struct<type: string>, operation: struct<type: string, enum: list<item: string>>, blend_factor_k: struct<type: string>, primitive_a: struct<type: string>, primitive_b: struct<type: string>, displacement_scale: struct<type: string>, shader_type: struct<type: string, enum: list<item: string>>, target: struct<type: string, enum: list<item: string>>, includes: struct<type: string, items: struct<type: string>>, output_format: struct<type: string, enum: list<item: string>>>, required: list<item: string>, additionalProperties: bool>>>>, tool_calls: list<item: struct<id: string, type: string, function: struct<name: string, arguments: string>>>>
              to
              {'role': Value('string'), 'content': Value('string')}
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1369, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      builder, max_dataset_size_bytes=max_dataset_size_bytes
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 948, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                  ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1683, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ~~~~~~~~~~~~~~~~~~~~~~~~~~^
                      gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  ):
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1869, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

messages list	images list	model_target string
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0000.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0001.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0002.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0003.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0004.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0005.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0006.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0007.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0008.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0009.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0010.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0011.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0012.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0013.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0014.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0015.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0016.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0017.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0018.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0019.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0020.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0021.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0022.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0023.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0024.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0025.png" ]	gemma_4_12b
[ { "role": "system", "content": "You are a specialized VLM for procedural terrain generation, shaders, and lighting (Inigo Quilez techniques).\nAvailable tools:\n[\n {\n \"name\": \"render_terrain_shader\",\n \"description\": \"Tool for render_terrain_shader\",\n \"parameters\": {\n \"type\": ...	[ "/Users/true/Documents/graphics/iq_terrain_vlm_dataset/dataset/images/code_gen_0026.png" ]	gemma_4_12b

End of preview.

IQ Terrain VLM Dataset

A high-fidelity, mathematically pristine Vision-Language Model (VLM) dataset designed specifically to teach models the procedural graphics and raymarching techniques of Inigo Quilez.

Dataset Summary

Most coding datasets rely on broadly scraped, often buggy code from GitHub or StackOverflow. This dataset takes a highly targeted approach:

Mathematical Ground Truth: All GLSL code and mathematical concepts are sourced directly from Inigo Quilez's foundational articles on procedural generation, SDFs (Signed Distance Fields), Fractional Brownian Motion (FBM), and raymarching.
Airtight Visual-Code Coupling: The dataset includes 1,053 procedurally generated ground-truth images. These images were mathematically generated to visually represent the exact mathematical concepts and shaders being taught in the text, providing an exceptionally tight visual-text pairing.
Cognitive Diversity: Rather than relying solely on "write code" prompts, the dataset is distributed across 6 distinct cognitive tasks to prevent memorization and teach structural understanding:
- Code Generation: Writing shaders from scratch based on a visual target.
- Code Explanation: Deconstructing and explaining complex raymarching loops.
- Code Debugging: Identifying and fixing logical errors in SDF or lighting math.
- Parameter Tuning: Adjusting constants (e.g., FBM octaves, lacunarity, fog density) to achieve specific visual results.
- Tool Calling: Outputting structured JSON to interface with external rendering pipelines.
- Multi-Turn Chat: Simulating a conversational debugging session.
Token Purity: The dataset has been rigorously scrubbed of HTML artifacts, rogue newlines, and colliding XML tags to ensure perfect integration with native VLM token templates (e.g., <|channel>thought>).

Dataset Structure

The dataset contains 1,002 total samples and is pre-formatted for two distinct VLM token templates:

Gemma 4 12B: Located in gemma4/train.jsonl and gemma4/val.jsonl
Step 3.7 Flash: Located in step37_flash/train.jsonl and step37_flash/val.jsonl

A paired images/ directory contains all 1,053 procedurally generated reference images (256x144 resolution) referenced by the JSONL files.

Training Recommendations (Avoiding Catastrophic Forgetting)

Because this dataset is 100% pure procedural graphics logic, fine-tuning a model exclusively on this dataset for many epochs will induce catastrophic forgetting of the model's base conversational abilities.

For a production-ready model, it is highly recommended to mix this dataset (e.g., at a 20% ratio) with a general instruction-tuning dataset (like UltraChat or ShareGPT, at an 80% ratio). This ensures the model learns the highly-specialized GLSL skills without losing its ability to hold a normal conversation.

License and Attribution

This dataset is released under the Creative Commons Attribution 4.0 International (CC BY 4.0) license.

You are free to share and adapt this dataset for any purpose, even commercially, under the condition that you provide appropriate credit and attribution to this original dataset repository.

Sourced Material: All underlying mathematical concepts and GLSL implementations within this dataset are derived from the public educational articles and techniques of Inigo Quilez.

Downloads last month: 66

Total file size:

50.6 MB