File size: 3,082 Bytes
cfced9a fee9962 cfced9a 85b63e5 fee9962 157b0f3 fee9962 157b0f3 fee9962 cfced9a fee9962 157b0f3 fee9962 157b0f3 fee9962 85b63e5 fee9962 157b0f3 85b63e5 fee9962 157b0f3 fee9962 157b0f3 fee9962 157b0f3 fee9962 157b0f3 fee9962 157b0f3 fee9962 157b0f3 fee9962 85b63e5 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 |
---
library_name: peft
base_model: cerebras/Cerebras-GPT-256M
license: apache-2.0
language:
- en
tags:
- pcgml
- pcg
- biome
- terrain
---
# Model Card for Model ID
Part of bachelor thesis focusing on generating 3d terrain from text.<br>
Biome Definition is a json formatted string detailing the biomes and their features. <br>
Together with SD Heightmap Generation v3.0, 3d terrain can be procedurally generated.
<!-- Provide a quick summary of what the model is/does. -->
## Uses
Example Inference:<br>
<prompt> Alpine mountains and a forest adorned with red flowers.
Result:<br>
<prompt> Alpine mountains and a forest adorned with red flowers.<br>
<biomes> [{'Entities': {'Boulders': [], 'BouldersSpawnrate': 0.0, 'Grass': [{'BaseTint': '#ff0000', 'Id': 'flowers'}], 'GrassSpawnrate': 0.8, 'Trees': [], 'TreesSpawnrate': 0.0}, 'Name':'Alpine Mountains', 'SpawnCondition': {'HeightRange': [0.0, 0.5], 'SlopeRange': [0.0, 0.3]}, 'Texturing': {'GroundTexture': 'grass', 'GroundTextureTint': '#7cfc00', 'SlopeTexture': 'grass', 'SlopeTextureTint': '#7cfc00'}}, {'Entities': {'Boulders': [], 'BouldersSpawnrate': 0.0, 'Grass': [], 'GrassSpawnrate': 0.0, 'Trees': [{'BaseTint': '#8b4513', 'Id': 'tree_large'}], 'TreesSpawnrate': 0.8}, 'Name':'Redwood Forest', 'SpawnCondition': {'HeightRange': [0.0, 0.5], 'SlopeRange': [0.0, 0.3]}, 'Texturing': {'GroundTexture': 'dirt', 'GroundTextureTint': '#7cfc00', 'SlopeTexture':'stone', 'SlopeTextureTint': '#c2c2c2'}}]
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
Requires ~3GB VRAM
Training Loss: 4%
Validation Loss: 6%
## Bias, Risks, and Limitations
Model is biased to certain ids like:
* textures: grass, dirt, mud, mud_dry, stone, stone_smooth
* boulders: boulder_small, boulder_medium, boulder_large
* trees: tree_small, tree_medium, tree_large, tree_conifer_small, ...
* grass: grass_small, grass_medium, grass_large, grass_dry, flowers
Model also has trouble overadjusting tints, leading to oversaturated colorations.
Model sometimes continues output. Just end prompt at "}}]".
Model sometimes creates incomplete biome definitions or places entities on wrong biome.
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
## Model Details
Finetuned on custom dataset of size 538 json formatted biome definitions consisting of:
* Name
* SpawnCondition
* HeightRange
* SlopeRange
* Entities
* Trees
* Id
* BaseTint
* Boulders
* Id
* BaseTint
* Grass
* Id
* BaseTint
* TreesSpawnrate
* BouldersSpawnrate
* GrassSpawnrate
* Texturing
* GroundTexture
* GroundTextureTint
* SlopeTexture
* SlopeTextureTint
### Model Description
<!-- Provide a longer summary of what this model is. -->
- **Model type:** [PEFT Adapter]
- **Language(s) (NLP):** [English]
- **License:** [Apache 2.0]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Paper [optional]:** [tba]
### Framework versions
- PEFT 0.11.0 |