Downloads · 30 days
33
4% of all-time downloads
canine-s-squad
Splend1dchan/canine-s-squad
canine-s-squad is a question answering model from Splend1dchan. Use it when the input is a question plus a passage. It is set up for transformers.
python runsquad.py --modelnameorpath google/canine-s --dotrain --doeval --pergputrainbatchsize 1 --pergpuevalbatchsize 1 --gradientaccumulationsteps 128 --learningrate 3e-5 --numtrainepochs 3 --maxseqlength 1024 --doc…
All-time downloads
771
Public
Repo size
1.1 GB
Likes
1
Public
Repo makeup
Click a slice to open those files.
.bin529 MB · 91%
At a glance
- Task
- Question Answering
- Library
- transformers
- Model type
- canine
- Access
- Public
- Created
- Apr 7, 2022
- Updated
- Apr 12, 2022
- SHA
- f59594f9
Try a prompt
Model card
From the Hugging Face model README
python run_squad.py --model_name_or_path google/canine-s --do_train --do_eval --per_gpu_train_batch_size 1 --per_gpu_eval_batch_size 1 --gradient_accumulation_steps 128 --learning_rate 3e-5 --num_train_epochs 3 --max_seq_length 1024 --doc_stride 128 --max_answer_length 240 --output_dir canine-s-squad --model_type bert
{ "_name_or_path": "google/canine-s", "architectures": [ "CanineForQuestionAnswering" ], "attention_probs_dropout_prob": 0.1, "bos_token_id": 57344, "downsampling_rate": 4, "eos_token_id": 57345, "hidden_act": "gelu", "hidden_dropout_prob": 0.1, "hidden_size": 768, "initializer_range": 0.02, "intermediate_size": 3072, "layer_norm_eps": 1e-12, "local_transformer_stride": 128, "max_position_embeddings": 16384, "model_type": "canine", "num_attention_heads": 12, "num_hash_buckets": 16384, "num_hash_functions": 8, "num_hidden_layers": 12, "pad_token_id": 0, "torch_dtype": "float32", "transformers_version": "4.19.0.dev0", "type_vocab_size": 16, "upsampling_kernel_size": 4, "use_cache": true }
{'exact': 64.70198675496688, 'f1': 76.57594921776277}
- Task
- Question Answering
- Library
- transformers
- Type
- canine
- Created
- Apr 7, 2022
- Updated
- Apr 12, 2022