Update config.json
Browse files- config.json +42 -25
config.json
CHANGED
@@ -1,35 +1,51 @@
|
|
1 |
{
|
2 |
-
"activation_function": "gelu_new",
|
3 |
-
"architectures": ["GPT2LMHeadModel"],
|
4 |
-
"attn_pdrop": 0.1,
|
5 |
-
"bos_token_id": 50256,
|
6 |
-
"embd_pdrop": 0.1,
|
7 |
-
"eos_token_id": 50256,
|
8 |
-
"initializer_range": 0.02,
|
9 |
-
"layer_norm_epsilon": 1e-05,
|
10 |
"model_type": "gpt2",
|
11 |
-
"
|
12 |
-
"
|
13 |
-
|
14 |
-
|
15 |
-
|
16 |
-
|
17 |
-
"
|
18 |
-
|
19 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
20 |
"task_specific_params": {
|
21 |
-
"
|
22 |
-
"
|
23 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
24 |
}
|
25 |
},
|
|
|
26 |
"language": ["en"],
|
27 |
-
"tags": ["
|
28 |
-
"metrics": ["accuracy"],
|
29 |
-
"pipeline_tag": "
|
30 |
"library_name": "transformers",
|
|
|
31 |
"license": "apache-2.0",
|
32 |
-
"is_quantized": true,
|
33 |
"custom_params": {
|
34 |
"adaptation_rate": 0.05,
|
35 |
"complexity_metric": null,
|
@@ -43,10 +59,11 @@
|
|
43 |
"initial_neuron_count": 5000,
|
44 |
"innovative_growth_net": {
|
45 |
"adaptation_rate": 0.05,
|
|
|
46 |
"initial_capacity": 250000,
|
47 |
"input_size": 2048
|
48 |
},
|
49 |
-
"input_dimension":
|
50 |
"low_stability_threshold": 0.01,
|
51 |
"max_complexity": 10000,
|
52 |
"max_neurons": 250000,
|
|
|
1 |
{
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2 |
"model_type": "gpt2",
|
3 |
+
"architectures": ["GPT2LMHeadModel"],
|
4 |
+
"tokenizer_config": {
|
5 |
+
"bos_token_id": 50256,
|
6 |
+
"eos_token_id": 50256,
|
7 |
+
"n_positions": 2048
|
8 |
+
},
|
9 |
+
"config": {
|
10 |
+
"activation_function": "gelu_new",
|
11 |
+
"n_ctx": 2048,
|
12 |
+
"n_embd": 2048,
|
13 |
+
"n_head": 16,
|
14 |
+
"n_layer": 24,
|
15 |
+
"n_positions": 2048,
|
16 |
+
"n_special": 0,
|
17 |
+
"attn_pdrop": 0.1,
|
18 |
+
"embd_pdrop": 0.1,
|
19 |
+
"initializer_range": 0.02,
|
20 |
+
"layer_norm_epsilon": 1e-05,
|
21 |
+
"resid_pdrop": 0.1,
|
22 |
+
"summary_activation": null,
|
23 |
+
"summary_first_dropout": 0.1,
|
24 |
+
"summary_proj_to_labels": true,
|
25 |
+
"summary_type": "cls_index",
|
26 |
+
"summary_use_proj": true
|
27 |
+
},
|
28 |
"task_specific_params": {
|
29 |
+
"conversational": {
|
30 |
+
"max_length": 1024,
|
31 |
+
"min_length": 20,
|
32 |
+
"length_penalty": 1.5,
|
33 |
+
"num_beams": 5,
|
34 |
+
"early_stopping": true,
|
35 |
+
"no_repeat_ngram_size": 3,
|
36 |
+
"temperature": 0.7,
|
37 |
+
"top_k": 50,
|
38 |
+
"top_p": 0.9
|
39 |
}
|
40 |
},
|
41 |
+
"transformers_version": "4.34.0",
|
42 |
"language": ["en"],
|
43 |
+
"tags": ["conversational"],
|
44 |
+
"metrics": ["perplexity", "accuracy"],
|
45 |
+
"pipeline_tag": "conversational",
|
46 |
"library_name": "transformers",
|
47 |
+
"datasets": ["vicgalle/alpaca-gpt4"],
|
48 |
"license": "apache-2.0",
|
|
|
49 |
"custom_params": {
|
50 |
"adaptation_rate": 0.05,
|
51 |
"complexity_metric": null,
|
|
|
59 |
"initial_neuron_count": 5000,
|
60 |
"innovative_growth_net": {
|
61 |
"adaptation_rate": 0.05,
|
62 |
+
"complexity_metric": null,
|
63 |
"initial_capacity": 250000,
|
64 |
"input_size": 2048
|
65 |
},
|
66 |
+
"input_dimension": 768,
|
67 |
"low_stability_threshold": 0.01,
|
68 |
"max_complexity": 10000,
|
69 |
"max_neurons": 250000,
|