Mediocreatmybest
commited on
Commit
•
6e83ea8
1
Parent(s):
89fa4f5
Upload InstructBlipForConditionalGeneration
Browse files- config.json +98 -0
- pytorch_model.bin +3 -0
config.json
ADDED
@@ -0,0 +1,98 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "Salesforce/instructblip-flan-t5-xl",
|
3 |
+
"architectures": [
|
4 |
+
"InstructBlipForConditionalGeneration"
|
5 |
+
],
|
6 |
+
"initializer_factor": 1.0,
|
7 |
+
"initializer_range": 0.02,
|
8 |
+
"is_encoder_decoder": true,
|
9 |
+
"model_type": "instructblip",
|
10 |
+
"num_query_tokens": 32,
|
11 |
+
"qformer_config": {
|
12 |
+
"classifier_dropout": null,
|
13 |
+
"model_type": "instructblip_qformer",
|
14 |
+
"vocab_size": 30523
|
15 |
+
},
|
16 |
+
"quantization_config": {
|
17 |
+
"bnb_4bit_compute_dtype": "bfloat16",
|
18 |
+
"bnb_4bit_quant_type": "nf4",
|
19 |
+
"bnb_4bit_use_double_quant": true,
|
20 |
+
"llm_int8_enable_fp32_cpu_offload": true,
|
21 |
+
"llm_int8_has_fp16_weight": false,
|
22 |
+
"llm_int8_skip_modules": null,
|
23 |
+
"llm_int8_threshold": 6.0,
|
24 |
+
"load_in_4bit": false,
|
25 |
+
"load_in_8bit": true,
|
26 |
+
"quant_method": "bitsandbytes"
|
27 |
+
},
|
28 |
+
"text_config": {
|
29 |
+
"architectures": [
|
30 |
+
"T5ForConditionalGeneration"
|
31 |
+
],
|
32 |
+
"bos_token_id": 1,
|
33 |
+
"classifier_dropout": 0.0,
|
34 |
+
"d_ff": 5120,
|
35 |
+
"d_kv": 64,
|
36 |
+
"d_model": 2048,
|
37 |
+
"decoder_start_token_id": 0,
|
38 |
+
"dense_act_fn": "gelu",
|
39 |
+
"dropout_rate": 0.1,
|
40 |
+
"eos_token_id": 1,
|
41 |
+
"feed_forward_proj": "gated-gelu",
|
42 |
+
"initializer_factor": 1.0,
|
43 |
+
"is_encoder_decoder": true,
|
44 |
+
"is_gated_act": true,
|
45 |
+
"layer_norm_epsilon": 1e-06,
|
46 |
+
"model_type": "t5",
|
47 |
+
"n_positions": 512,
|
48 |
+
"num_decoder_layers": 24,
|
49 |
+
"num_heads": 32,
|
50 |
+
"num_layers": 24,
|
51 |
+
"output_past": true,
|
52 |
+
"pad_token_id": 0,
|
53 |
+
"relative_attention_max_distance": 128,
|
54 |
+
"relative_attention_num_buckets": 32,
|
55 |
+
"task_specific_params": {
|
56 |
+
"summarization": {
|
57 |
+
"early_stopping": true,
|
58 |
+
"length_penalty": 2.0,
|
59 |
+
"max_length": 200,
|
60 |
+
"min_length": 30,
|
61 |
+
"no_repeat_ngram_size": 3,
|
62 |
+
"num_beams": 4,
|
63 |
+
"prefix": "summarize: "
|
64 |
+
},
|
65 |
+
"translation_en_to_de": {
|
66 |
+
"early_stopping": true,
|
67 |
+
"max_length": 300,
|
68 |
+
"num_beams": 4,
|
69 |
+
"prefix": "translate English to German: "
|
70 |
+
},
|
71 |
+
"translation_en_to_fr": {
|
72 |
+
"early_stopping": true,
|
73 |
+
"max_length": 300,
|
74 |
+
"num_beams": 4,
|
75 |
+
"prefix": "translate English to French: "
|
76 |
+
},
|
77 |
+
"translation_en_to_ro": {
|
78 |
+
"early_stopping": true,
|
79 |
+
"max_length": 300,
|
80 |
+
"num_beams": 4,
|
81 |
+
"prefix": "translate English to Romanian: "
|
82 |
+
}
|
83 |
+
},
|
84 |
+
"tie_word_embeddings": false,
|
85 |
+
"torch_dtype": "float32",
|
86 |
+
"vocab_size": 32128
|
87 |
+
},
|
88 |
+
"tie_word_embeddings": false,
|
89 |
+
"torch_dtype": "float16",
|
90 |
+
"transformers_version": "4.32.0.dev0",
|
91 |
+
"use_decoder_only_language_model": false,
|
92 |
+
"vision_config": {
|
93 |
+
"dropout": 0.0,
|
94 |
+
"initializer_factor": 1.0,
|
95 |
+
"model_type": "instructblip_vision_model",
|
96 |
+
"num_channels": 3
|
97 |
+
}
|
98 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:946a27f67b563504dd1a08b06940b32f6f79f0bc5be11567235cac1b38fd0dd3
|
3 |
+
size 5745359505
|