Upload model
Browse files- config.json +6 -1
- modeling_mamba.py +6 -3
config.json
CHANGED
|
@@ -1,6 +1,10 @@
|
|
| 1 |
{
|
|
|
|
|
|
|
|
|
|
| 2 |
"auto_map": {
|
| 3 |
-
"AutoConfig": "configuration_mamba.MambaConfig"
|
|
|
|
| 4 |
},
|
| 5 |
"bias": false,
|
| 6 |
"conv_bias": true,
|
|
@@ -14,6 +18,7 @@
|
|
| 14 |
"model_type": "mamba",
|
| 15 |
"n_layer": 24,
|
| 16 |
"pad_vocab_size_multiple": 8,
|
|
|
|
| 17 |
"transformers_version": "4.37.2",
|
| 18 |
"vocab_size": 50280
|
| 19 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"MambaModelForCausalLM"
|
| 4 |
+
],
|
| 5 |
"auto_map": {
|
| 6 |
+
"AutoConfig": "configuration_mamba.MambaConfig",
|
| 7 |
+
"AutoModelForCausalLM": "modeling_mamba.MambaModelForCausalLM"
|
| 8 |
},
|
| 9 |
"bias": false,
|
| 10 |
"conv_bias": true,
|
|
|
|
| 18 |
"model_type": "mamba",
|
| 19 |
"n_layer": 24,
|
| 20 |
"pad_vocab_size_multiple": 8,
|
| 21 |
+
"torch_dtype": "float32",
|
| 22 |
"transformers_version": "4.37.2",
|
| 23 |
"vocab_size": 50280
|
| 24 |
}
|
modeling_mamba.py
CHANGED
|
@@ -434,9 +434,12 @@ class MambaModelForCausalLM(MambaPreTrainedModel):
|
|
| 434 |
loss=loss,
|
| 435 |
)
|
| 436 |
|
| 437 |
-
def prepare_inputs_for_generation(
|
| 438 |
-
|
| 439 |
-
|
|
|
|
|
|
|
|
|
|
| 440 |
|
| 441 |
|
| 442 |
class MambaModelForSequenceClassification(MambaPreTrainedModel):
|
|
|
|
| 434 |
loss=loss,
|
| 435 |
)
|
| 436 |
|
| 437 |
+
def prepare_inputs_for_generation(
|
| 438 |
+
self, input_ids, attention_mask=None, **model_kwargs
|
| 439 |
+
):
|
| 440 |
+
return {
|
| 441 |
+
"input_ids": input_ids,
|
| 442 |
+
}
|
| 443 |
|
| 444 |
|
| 445 |
class MambaModelForSequenceClassification(MambaPreTrainedModel):
|