tarekziade HF Staff
Publish MIC usage, slim fixture and reviewed Transformers expectations
ff3c8d1 verified | { | |
| "fixture": "huggingface-mic/roberta-base-mic", | |
| "key": "05ec051e3d97c7210a49a8776fb18ac2156a51f1e08fedd0de7d1af05e8368f0", | |
| "key_fields": { | |
| "architectures": [ | |
| "RobertaForMaskedLM" | |
| ], | |
| "output_hidden_states": false, | |
| "return_dict": true, | |
| "chunk_size_feed_forward": 0, | |
| "is_encoder_decoder": false, | |
| "id2label": { | |
| "0": "LABEL_0", | |
| "1": "LABEL_1" | |
| }, | |
| "label2id": { | |
| "LABEL_0": 0, | |
| "LABEL_1": 1 | |
| }, | |
| "problem_type": null, | |
| "vocab_size": 50265, | |
| "hidden_size": 64, | |
| "num_hidden_layers": 2, | |
| "num_attention_heads": 4, | |
| "intermediate_size": 128, | |
| "hidden_act": "gelu", | |
| "hidden_dropout_prob": 0.1, | |
| "attention_probs_dropout_prob": 0.1, | |
| "max_position_embeddings": 514, | |
| "type_vocab_size": 1, | |
| "initializer_range": 0.02, | |
| "layer_norm_eps": 1e-05, | |
| "pad_token_id": 1, | |
| "bos_token_id": 0, | |
| "eos_token_id": 2, | |
| "use_cache": true, | |
| "classifier_dropout": null, | |
| "is_decoder": false, | |
| "add_cross_attention": false, | |
| "tie_word_embeddings": true, | |
| "model_type": "roberta", | |
| "output_attentions": false | |
| }, | |
| "covers": { | |
| "FacebookAI/roberta-base": "e2da8e2f811d1448a5b465c236feacd80ffbac7b" | |
| }, | |
| "built_from": { | |
| "repo": "FacebookAI/roberta-base", | |
| "revision": "e2da8e2f811d1448a5b465c236feacd80ffbac7b" | |
| }, | |
| "seed": 17, | |
| "dtype": "float32", | |
| "builder": "example/build_fixtures.py", | |
| "model_class": "RobertaForMaskedLM", | |
| "parameters": 3371545, | |
| "missing_parameters": [], | |
| "checkpoint_inspection": "parameter headers/indexes; bounded binary fallback", | |
| "numeric_health_scope": "CPU float32/bfloat16/float16; autocast for low-precision losses", | |
| "captured": { | |
| "call": "training step", | |
| "inputs": [ | |
| "attention_mask", | |
| "input_ids", | |
| "labels" | |
| ], | |
| "class": "RobertaForMaskedLM" | |
| }, | |
| "weights": "random", | |
| "gate": { | |
| "random": { | |
| "pass": true, | |
| "accepted_insensitive": {}, | |
| "threshold": 0.001, | |
| "deltas": { | |
| "input_ids:roll_batch": 0.6873351335525513, | |
| "input_ids:reverse": 0.8940852284431458, | |
| "attention_mask:all_ones": 0.002395316958427429 | |
| }, | |
| "weakest": "attention_mask:all_ones", | |
| "no_perturbation_defined": [] | |
| } | |
| }, | |
| "numeric_health": { | |
| "float32": { | |
| "outputs_finite": true, | |
| "loss_finite": true, | |
| "gradients_finite": true, | |
| "finite": true | |
| }, | |
| "bfloat16": { | |
| "outputs_finite": true, | |
| "loss_finite": true, | |
| "gradients_finite": true, | |
| "finite": true | |
| }, | |
| "float16": { | |
| "outputs_finite": true, | |
| "loss_finite": true, | |
| "unscaled_gradients_finite": true, | |
| "finite": true | |
| } | |
| }, | |
| "copied_unchanged": [ | |
| "dict.txt", | |
| "merges.txt", | |
| "tokenizer.json", | |
| "tokenizer_config.json", | |
| "vocab.json" | |
| ], | |
| "environment": { | |
| "python": "3.12.12", | |
| "machine": "arm64", | |
| "torch": "2.14.1", | |
| "transformers": "5.22.0.dev0", | |
| "tokenizers": "0.23.3" | |
| }, | |
| "sha256": { | |
| "config.json": "830930b5ae6b215a3ec3b24228c61179c7965a1a519bb3abed82630a6c3ad06c", | |
| "dict.txt": "2c31c5b236661067fe8626775eb995ef5fe808ac87d2ad4a6ddbef68550f5cc6", | |
| "merges.txt": "1ce1664773c50f3e0cc8842619a93edc4624525b728b188a9e0be33b7726adc5", | |
| "model.safetensors": "039fe2a956ec50b3d5fe667490873e9d1697b906fa0d9f82cf5a727c53c97bbf", | |
| "tokenizer.json": "847bbeab6174d66a88898f729d52fa8d355fafe1bea101cf960dd404581df70e", | |
| "tokenizer_config.json": "994f46754c5bf4014f1aa92d34b1374319c3a6b3f702105cd5b742beaecd18ce", | |
| "vocab.json": "9e7f63c2d15d666b52e21d250d2e513b87c9b713cfa6987a82ed89e5e6e50655" | |
| } | |
| } | |