Instructions to use Mishamq/HybriDNA-3B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Mishamq/HybriDNA-3B with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("Mishamq/HybriDNA-3B", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Fix import path for trust_remote_code loading
Browse filesReplace `from hf.configuration_hybridna` with relative import
`from .configuration_hybridna` so the model loads correctly via
AutoModel.from_pretrained(..., trust_remote_code=True).
- modeling_hybridna.py +1 -1
modeling_hybridna.py
CHANGED
|
@@ -53,7 +53,7 @@ from transformers.utils.import_utils import (
|
|
| 53 |
is_flash_attn_2_available,
|
| 54 |
is_mamba_ssm_available,
|
| 55 |
)
|
| 56 |
-
from
|
| 57 |
from mamba_ssm.ops.triton.layernorm_gated import RMSNorm as RMSNormGated
|
| 58 |
|
| 59 |
# try except block so it'll work with trust_remote_code.
|
|
|
|
| 53 |
is_flash_attn_2_available,
|
| 54 |
is_mamba_ssm_available,
|
| 55 |
)
|
| 56 |
+
from .configuration_hybridna import HybriDNAConfig
|
| 57 |
from mamba_ssm.ops.triton.layernorm_gated import RMSNorm as RMSNormGated
|
| 58 |
|
| 59 |
# try except block so it'll work with trust_remote_code.
|