File size: 3,107 Bytes
deeef7d 9bce8c9 ad4998e 9bce8c9 ad4998e 9bce8c9 ad4998e 9bce8c9 deeef7d 9bce8c9 ad4998e 9bce8c9 deeef7d ad4998e 9bce8c9 ad4998e 9bce8c9 ad4998e 9bce8c9 ad4998e 9bce8c9 ad4998e 9bce8c9 ad4998e 9bce8c9 ad4998e 9bce8c9 ad4998e 9bce8c9 ad4998e 9bce8c9 |
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 91 |
---
library_name: peft
base_model: tiiuae/falcon-rw-1b
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': 0.0, 'Trees': [], 'TreesSpawnrate': 0.0}, 'Name': 'Red Flowers', 'SpawnCondition': {'HeightRange': [0.0, 0.1], 'SlopeRange': [0.0, 0.1]}, 'Texturing': {'GroundTexture': 'grass', 'GroundTextureTint': '#7cfc00', 'SlopeTexture': 'grass', 'SlopeTextureTint': '#7cfc00'}}, {'Entities': {'Boulders': [], 'BouldersSpawnrate': 0.0, 'Grass': [], 'GrassSpawnrate': 0.0, 'Trees': [{'BaseTint': '#ffffff', 'Id': 'tree_medium'}], 'TreesSpawnrate': 0.8}, 'Name': 'Mountain Range', 'SpawnCondition': {'HeightRange': [0.0, 0.7], 'SlopeRange': [0.0, 0.5]}, 'Texturing': {'GroundTexture':'stone', 'GroundTextureTint': '#ffffff', 'SlopeTexture':'stone_smooth', 'SlopeTextureTint': '#ffffff'}}]
<!-- 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: ~8GB VRAM
Training Loss: 2%
Validation Loss: 4%
## 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]
[More Information Needed]
### Framework versions
- PEFT 0.11.0 |