Upload XomdichForConditionalGeneration
Browse files- config.json +12 -7
- generation_config.json +1 -1
- pytorch_model.bin +2 -2
config.json
CHANGED
|
@@ -4,16 +4,21 @@
|
|
| 4 |
],
|
| 5 |
"attention_dropout": 0.1,
|
| 6 |
"bos_token_id": 1,
|
|
|
|
|
|
|
|
|
|
| 7 |
"eos_token_id": 2,
|
| 8 |
-
"
|
|
|
|
| 9 |
"max_sequence_length": 512,
|
| 10 |
-
"model_type": "
|
| 11 |
-
"num_attention_heads":
|
| 12 |
-
"num_key_value_heads":
|
| 13 |
-
"num_layers":
|
| 14 |
-
"pad_token_id":
|
|
|
|
| 15 |
"torch_dtype": "float32",
|
| 16 |
-
"transformers_version": "4.
|
| 17 |
"use_cache": true,
|
| 18 |
"vocab_size": 25001
|
| 19 |
}
|
|
|
|
| 4 |
],
|
| 5 |
"attention_dropout": 0.1,
|
| 6 |
"bos_token_id": 1,
|
| 7 |
+
"d_ff": 3072,
|
| 8 |
+
"d_kv": 192,
|
| 9 |
+
"d_model": 768,
|
| 10 |
"eos_token_id": 2,
|
| 11 |
+
"feed_forward_proj": "relu",
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
"max_sequence_length": 512,
|
| 14 |
+
"model_type": "hyper-xomdich",
|
| 15 |
+
"num_attention_heads": 4,
|
| 16 |
+
"num_key_value_heads": 1,
|
| 17 |
+
"num_layers": 26,
|
| 18 |
+
"pad_token_id": 25000,
|
| 19 |
+
"relative_attention_num_buckets": 32,
|
| 20 |
"torch_dtype": "float32",
|
| 21 |
+
"transformers_version": "4.50.3",
|
| 22 |
"use_cache": true,
|
| 23 |
"vocab_size": 25001
|
| 24 |
}
|
generation_config.json
CHANGED
|
@@ -3,5 +3,5 @@
|
|
| 3 |
"bos_token_id": 1,
|
| 4 |
"eos_token_id": 2,
|
| 5 |
"pad_token_id": 0,
|
| 6 |
-
"transformers_version": "4.
|
| 7 |
}
|
|
|
|
| 3 |
"bos_token_id": 1,
|
| 4 |
"eos_token_id": 2,
|
| 5 |
"pad_token_id": 0,
|
| 6 |
+
"transformers_version": "4.50.3"
|
| 7 |
}
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e43d515e8792b4086347b0331ed9b45706000c2e6f2abab863344930fee3a068
|
| 3 |
+
size 1766669487
|