Upload Two-Tower model
Browse files- best_model.pt +3 -0
- config.yml +31 -0
best_model.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4bb9033e31ac60f65170f8a07e935b2f36f010b9730638a6f7c51d451f0ebb3b
|
| 3 |
+
size 566119
|
config.yml
ADDED
|
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
batch_size: 256
|
| 2 |
+
checkpoint_dir: checkpoints
|
| 3 |
+
data: data/processed/classic_triplets.parquet
|
| 4 |
+
device: cpu
|
| 5 |
+
embedding:
|
| 6 |
+
embedding_dim: 64
|
| 7 |
+
type: lookup
|
| 8 |
+
encoder:
|
| 9 |
+
arch: mean
|
| 10 |
+
hidden_dim: 128
|
| 11 |
+
tied_weights: true
|
| 12 |
+
epochs: 3
|
| 13 |
+
huggingface:
|
| 14 |
+
private: false
|
| 15 |
+
push_to_hub: true
|
| 16 |
+
repo_id: mlx7-two-tower-retrieval
|
| 17 |
+
learning_rate: 1e-3
|
| 18 |
+
loss:
|
| 19 |
+
margin: 0.2
|
| 20 |
+
type: triplet
|
| 21 |
+
max_sequence_length: 64
|
| 22 |
+
optimizer:
|
| 23 |
+
lr: 0.001
|
| 24 |
+
type: adamw
|
| 25 |
+
tokeniser:
|
| 26 |
+
max_len: 64
|
| 27 |
+
type: char
|
| 28 |
+
use_wandb: true
|
| 29 |
+
wandb:
|
| 30 |
+
entity: azuremis
|
| 31 |
+
project: two-tower-retrieval
|