Upload text_encoder_2/config.json with huggingface_hub
Browse files- text_encoder_2/config.json +57 -0
text_encoder_2/config.json
ADDED
|
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_attn_implementation_autoset": true,
|
| 3 |
+
"architectures": [
|
| 4 |
+
"T5EncoderModel"
|
| 5 |
+
],
|
| 6 |
+
"classifier_dropout": 0.0,
|
| 7 |
+
"d_ff": 10240,
|
| 8 |
+
"d_kv": 64,
|
| 9 |
+
"d_model": 4096,
|
| 10 |
+
"decoder_start_token_id": 0,
|
| 11 |
+
"dense_act_fn": "gelu_new",
|
| 12 |
+
"dropout_rate": 0.1,
|
| 13 |
+
"eos_token_id": 1,
|
| 14 |
+
"export_model_type": "t5-encoder",
|
| 15 |
+
"feed_forward_proj": "gated-gelu",
|
| 16 |
+
"initializer_factor": 1.0,
|
| 17 |
+
"is_encoder_decoder": true,
|
| 18 |
+
"is_gated_act": true,
|
| 19 |
+
"layer_norm_epsilon": 1e-06,
|
| 20 |
+
"model_type": "t5",
|
| 21 |
+
"neuron": {
|
| 22 |
+
"auto_cast": null,
|
| 23 |
+
"auto_cast_type": null,
|
| 24 |
+
"compiler_type": "neuronx-cc",
|
| 25 |
+
"compiler_version": "2.19.8089.0+8ab9f450",
|
| 26 |
+
"dynamic_batch_size": false,
|
| 27 |
+
"float_dtype": "fp32",
|
| 28 |
+
"inline_weights_to_neff": false,
|
| 29 |
+
"input_names": [
|
| 30 |
+
"input_ids"
|
| 31 |
+
],
|
| 32 |
+
"int_dtype": "int64",
|
| 33 |
+
"model_type": "t5-encoder",
|
| 34 |
+
"optlevel": "2",
|
| 35 |
+
"output_attentions": false,
|
| 36 |
+
"output_hidden_states": false,
|
| 37 |
+
"output_names": [
|
| 38 |
+
"last_hidden_state"
|
| 39 |
+
],
|
| 40 |
+
"static_batch_size": 1,
|
| 41 |
+
"static_sequence_length": 512,
|
| 42 |
+
"task": "feature-extraction",
|
| 43 |
+
"tensor_parallel_size": 8
|
| 44 |
+
},
|
| 45 |
+
"num_decoder_layers": 24,
|
| 46 |
+
"num_heads": 64,
|
| 47 |
+
"num_layers": 24,
|
| 48 |
+
"output_past": true,
|
| 49 |
+
"pad_token_id": 0,
|
| 50 |
+
"relative_attention_max_distance": 128,
|
| 51 |
+
"relative_attention_num_buckets": 32,
|
| 52 |
+
"tie_word_embeddings": false,
|
| 53 |
+
"torch_dtype": "bfloat16",
|
| 54 |
+
"transformers_version": "4.51.3",
|
| 55 |
+
"use_cache": true,
|
| 56 |
+
"vocab_size": 32128
|
| 57 |
+
}
|