tarekziade HF Staff
Consolidate MIC usage, slim weights and reviewed framework expectations
f392307 verified | { | |
| "fixture": "hf-internal-testing/mic-t5-slim", | |
| "key": "c5306fc184517367b7284eb6dadf3108aefc3e6a7a788afac8474637ddc09a5b", | |
| "key_fields": { | |
| "architectures": [ | |
| "T5ForConditionalGeneration" | |
| ], | |
| "output_hidden_states": false, | |
| "return_dict": true, | |
| "chunk_size_feed_forward": 0, | |
| "is_encoder_decoder": true, | |
| "id2label": { | |
| "0": "LABEL_0", | |
| "1": "LABEL_1" | |
| }, | |
| "label2id": { | |
| "LABEL_0": 0, | |
| "LABEL_1": 1 | |
| }, | |
| "problem_type": null, | |
| "vocab_size": 32128, | |
| "d_model": 64, | |
| "d_kv": 16, | |
| "d_ff": 128, | |
| "num_layers": 2, | |
| "num_decoder_layers": 2, | |
| "num_heads": 4, | |
| "relative_attention_num_buckets": 32, | |
| "relative_attention_max_distance": 128, | |
| "dropout_rate": 0.1, | |
| "layer_norm_epsilon": 1e-06, | |
| "initializer_factor": 1.0, | |
| "feed_forward_proj": "relu", | |
| "use_cache": true, | |
| "pad_token_id": 0, | |
| "eos_token_id": 1, | |
| "classifier_dropout": 0.0, | |
| "is_decoder": false, | |
| "dense_act_fn": "relu", | |
| "is_gated_act": false, | |
| "scale_decoder_outputs": true, | |
| "tie_word_embeddings": true, | |
| "decoder_start_token_id": 0, | |
| "model_type": "t5", | |
| "n_positions": 512, | |
| "output_past": true, | |
| "task_specific_params": { | |
| "summarization": { | |
| "early_stopping": true, | |
| "length_penalty": 2.0, | |
| "max_length": 200, | |
| "min_length": 30, | |
| "no_repeat_ngram_size": 3, | |
| "num_beams": 4, | |
| "prefix": "summarize: " | |
| }, | |
| "translation_en_to_de": { | |
| "early_stopping": true, | |
| "max_length": 300, | |
| "num_beams": 4, | |
| "prefix": "translate English to German: " | |
| }, | |
| "translation_en_to_fr": { | |
| "early_stopping": true, | |
| "max_length": 300, | |
| "num_beams": 4, | |
| "prefix": "translate English to French: " | |
| }, | |
| "translation_en_to_ro": { | |
| "early_stopping": true, | |
| "max_length": 300, | |
| "num_beams": 4, | |
| "prefix": "translate English to Romanian: " | |
| } | |
| }, | |
| "output_attentions": false | |
| }, | |
| "covers": { | |
| "google-t5/t5-small": "df1b051c49625cf57a3d0d8d3863ed4d13564fe4" | |
| }, | |
| "built_from": { | |
| "repo": "google-t5/t5-small", | |
| "revision": "df1b051c49625cf57a3d0d8d3863ed4d13564fe4" | |
| }, | |
| "seed": 17, | |
| "dtype": "float32", | |
| "builder": "example/build_fixtures.py", | |
| "model_class": "T5ForConditionalGeneration", | |
| "parameters": 2221056, | |
| "missing_parameters": [], | |
| "checkpoint_inspection": "parameter metadata", | |
| "captured": { | |
| "call": "training step", | |
| "inputs": [ | |
| "attention_mask", | |
| "input_ids", | |
| "labels" | |
| ] | |
| }, | |
| "weights": "random", | |
| "gate": { | |
| "random": { | |
| "pass": true, | |
| "accepted_insensitive": {}, | |
| "threshold": 0.001, | |
| "deltas": { | |
| "input_ids:roll_batch": 3.3316690921783447, | |
| "input_ids:reverse": 2.0183849334716797, | |
| "attention_mask:all_ones": 1.2822163105010986 | |
| }, | |
| "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, | |
| "gradients_finite": true, | |
| "finite": true | |
| } | |
| }, | |
| "copied_unchanged": [ | |
| "generation_config.json", | |
| "spiece.model", | |
| "tokenizer.json", | |
| "tokenizer_config.json" | |
| ], | |
| "environment": { | |
| "python": "3.10.12", | |
| "machine": "x86_64", | |
| "torch": "2.11.0+cu128", | |
| "transformers": "5.19.0.dev0", | |
| "tokenizers": "0.23.1" | |
| }, | |
| "sha256": { | |
| "config.json": "055fa9bf51263b4af3ab65af8d3e61f1d3062bb9c55fa3113405887b91ee5c26", | |
| "generation_config.json": "f5a1c7e2be8092018d8835128987edf0111637dd98e90599cc80310fef75d95a", | |
| "model.safetensors": "35e86de468e80c899e4617c834c9cee408ee2299257d0ab7c8692c6fc808cb2b", | |
| "spiece.model": "d60acb128cf7b7f2536e8f38a5b18a05535c9e14c7a355904270e15b0945ea86", | |
| "tokenizer.json": "d2acde0d8d71dd30a711834b07781b9c89feaac33fd332f60507699282740066", | |
| "tokenizer_config.json": "d1e7146101aa96282057f374aa9c3b260fba2109e5990edb1775d6efc70ffa3c" | |
| } | |
| } | |