Upload model
Browse files- config.json +4 -5
- model/config.json +2 -2
- model/generation_config.json +1 -1
- model/model-00001-of-00002.safetensors +1 -1
- model/model-00002-of-00002.safetensors +1 -1
- pointer_projection.bin +1 -1
config.json
CHANGED
|
@@ -1,6 +1,5 @@
|
|
| 1 |
{
|
| 2 |
-
"
|
| 3 |
-
"_name_or_path": "../checkpoints/comment_resolution/pointers/epoch_2",
|
| 4 |
"architectures": [
|
| 5 |
"LlamaForCausalLM"
|
| 6 |
],
|
|
@@ -15,7 +14,7 @@
|
|
| 15 |
"intermediate_size": 5504,
|
| 16 |
"max_position_embeddings": 16384,
|
| 17 |
"mlp_bias": false,
|
| 18 |
-
"model_path": "/tmp/
|
| 19 |
"num_attention_heads": 16,
|
| 20 |
"num_hidden_layers": 24,
|
| 21 |
"num_key_value_heads": 16,
|
|
@@ -28,9 +27,9 @@
|
|
| 28 |
},
|
| 29 |
"rope_theta": 100000,
|
| 30 |
"tie_word_embeddings": false,
|
| 31 |
-
"tokenizer_path": "/tmp/
|
| 32 |
"torch_dtype": "float32",
|
| 33 |
-
"transformers_version": "4.47.
|
| 34 |
"use_cache": true,
|
| 35 |
"vocab_size": 32027
|
| 36 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "../checkpoints/comment_resolution/training/model/deepseek-ai/deepseek-coder-1.3b-instruct_with_new_tokens_three_edits_fixed/epoch_1",
|
|
|
|
| 3 |
"architectures": [
|
| 4 |
"LlamaForCausalLM"
|
| 5 |
],
|
|
|
|
| 14 |
"intermediate_size": 5504,
|
| 15 |
"max_position_embeddings": 16384,
|
| 16 |
"mlp_bias": false,
|
| 17 |
+
"model_path": "/tmp/tmpvlynhjyy/model",
|
| 18 |
"num_attention_heads": 16,
|
| 19 |
"num_hidden_layers": 24,
|
| 20 |
"num_key_value_heads": 16,
|
|
|
|
| 27 |
},
|
| 28 |
"rope_theta": 100000,
|
| 29 |
"tie_word_embeddings": false,
|
| 30 |
+
"tokenizer_path": "/tmp/tmpvlynhjyy/tokenizer",
|
| 31 |
"torch_dtype": "float32",
|
| 32 |
+
"transformers_version": "4.47.1",
|
| 33 |
"use_cache": true,
|
| 34 |
"vocab_size": 32027
|
| 35 |
}
|
model/config.json
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
{
|
| 2 |
-
"_name_or_path": "../checkpoints/comment_resolution/
|
| 3 |
"architectures": [
|
| 4 |
"LlamaForCausalLM"
|
| 5 |
],
|
|
@@ -28,7 +28,7 @@
|
|
| 28 |
"rope_theta": 100000,
|
| 29 |
"tie_word_embeddings": false,
|
| 30 |
"torch_dtype": "float32",
|
| 31 |
-
"transformers_version": "4.47.
|
| 32 |
"use_cache": true,
|
| 33 |
"vocab_size": 32027
|
| 34 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "../checkpoints/comment_resolution/training/model/deepseek-ai/deepseek-coder-1.3b-instruct_with_new_tokens_three_edits_fixed/epoch_1",
|
| 3 |
"architectures": [
|
| 4 |
"LlamaForCausalLM"
|
| 5 |
],
|
|
|
|
| 28 |
"rope_theta": 100000,
|
| 29 |
"tie_word_embeddings": false,
|
| 30 |
"torch_dtype": "float32",
|
| 31 |
+
"transformers_version": "4.47.1",
|
| 32 |
"use_cache": true,
|
| 33 |
"vocab_size": 32027
|
| 34 |
}
|
model/generation_config.json
CHANGED
|
@@ -2,5 +2,5 @@
|
|
| 2 |
"_from_model_config": true,
|
| 3 |
"bos_token_id": 32013,
|
| 4 |
"eos_token_id": 32021,
|
| 5 |
-
"transformers_version": "4.47.
|
| 6 |
}
|
|
|
|
| 2 |
"_from_model_config": true,
|
| 3 |
"bos_token_id": 32013,
|
| 4 |
"eos_token_id": 32021,
|
| 5 |
+
"transformers_version": "4.47.1"
|
| 6 |
}
|
model/model-00001-of-00002.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4984504096
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eb946d583830f7c9e0553ebf705075fbbc9e2107afead3894d5bea0853deed92
|
| 3 |
size 4984504096
|
model/model-00002-of-00002.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 397656840
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:81979ac52e1b4f5b49637f761eb6ebe41d664d8d4ad73e2dc76abdd85b63dd31
|
| 3 |
size 397656840
|
pointer_projection.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 16786970
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dd54d7dcefec419c694337b906c43fdb4543369529d71fc568969b0fb8dc3b6b
|
| 3 |
size 16786970
|