Transformers
PyTorch
English
language-model
graph-attention
adaptive-depth
temporal-decay
efficient-llm
Eval Results (legacy)
Instructions to use vigneshwar234/TemporalMesh-Transformer with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use vigneshwar234/TemporalMesh-Transformer with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("vigneshwar234/TemporalMesh-Transformer", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Add DOI, expanded tags, full benchmark metrics to model card YAML
Browse files
README.md
CHANGED
|
@@ -17,12 +17,21 @@ tags:
|
|
| 17 |
- causal-lm
|
| 18 |
- research
|
| 19 |
- preprint
|
|
|
|
|
|
|
|
|
|
|
|
|
| 20 |
library_name: pytorch
|
| 21 |
pipeline_tag: text-generation
|
| 22 |
datasets:
|
| 23 |
- vigneshwar234/TMT-Benchmarks
|
| 24 |
metrics:
|
| 25 |
- perplexity
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 26 |
model-index:
|
| 27 |
- name: TemporalMesh Transformer (TMT-Base)
|
| 28 |
results:
|
|
@@ -32,11 +41,35 @@ model-index:
|
|
| 32 |
dataset:
|
| 33 |
type: wikitext
|
| 34 |
name: WikiText-2
|
|
|
|
|
|
|
| 35 |
metrics:
|
| 36 |
- type: perplexity
|
| 37 |
value: 29.4
|
| 38 |
name: Validation Perplexity
|
| 39 |
verified: false
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 40 |
---
|
| 41 |
|
| 42 |
<div align="center">
|
|
|
|
| 17 |
- causal-lm
|
| 18 |
- research
|
| 19 |
- preprint
|
| 20 |
+
- mesh-transformer
|
| 21 |
+
- dynamic-graph
|
| 22 |
+
- early-exit
|
| 23 |
+
- per-token-routing
|
| 24 |
library_name: pytorch
|
| 25 |
pipeline_tag: text-generation
|
| 26 |
datasets:
|
| 27 |
- vigneshwar234/TMT-Benchmarks
|
| 28 |
metrics:
|
| 29 |
- perplexity
|
| 30 |
+
doi: 10.5281/zenodo.20287390
|
| 31 |
+
extra_gated_prompt: |
|
| 32 |
+
Paper DOI: https://doi.org/10.5281/zenodo.20287390
|
| 33 |
+
Zenodo: https://zenodo.org/records/20287390
|
| 34 |
+
GitHub: https://github.com/vignesh2027/TemporalMesh-Transformer
|
| 35 |
model-index:
|
| 36 |
- name: TemporalMesh Transformer (TMT-Base)
|
| 37 |
results:
|
|
|
|
| 41 |
dataset:
|
| 42 |
type: wikitext
|
| 43 |
name: WikiText-2
|
| 44 |
+
config: wikitext-2-raw-v1
|
| 45 |
+
split: validation
|
| 46 |
metrics:
|
| 47 |
- type: perplexity
|
| 48 |
value: 29.4
|
| 49 |
name: Validation Perplexity
|
| 50 |
verified: false
|
| 51 |
+
- task:
|
| 52 |
+
type: text-generation
|
| 53 |
+
name: Efficient Inference
|
| 54 |
+
dataset:
|
| 55 |
+
type: wikitext
|
| 56 |
+
name: WikiText-2
|
| 57 |
+
config: wikitext-2-raw-v1
|
| 58 |
+
split: validation
|
| 59 |
+
metrics:
|
| 60 |
+
- type: perplexity
|
| 61 |
+
value: 29.4
|
| 62 |
+
name: Validation Perplexity
|
| 63 |
+
verified: false
|
| 64 |
+
- name: Relative Compute
|
| 65 |
+
type: efficiency
|
| 66 |
+
value: 0.48
|
| 67 |
+
verified: false
|
| 68 |
+
- name: Avg Exit Layer
|
| 69 |
+
type: efficiency
|
| 70 |
+
value: 5.5
|
| 71 |
+
verified: false
|
| 72 |
+
---
|
| 73 |
---
|
| 74 |
|
| 75 |
<div align="center">
|