{ "fixture": "huggingface-mic/phi-2-mic", "key": "97b6a16aa5670860288f384e8955b1ba3b0995d7c7124e3049ffd39659c0a7aa", "key_fields": { "architectures": [ "PhiForCausalLM" ], "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": 51200, "hidden_size": 64, "intermediate_size": 128, "num_hidden_layers": 2, "num_attention_heads": 4, "num_key_value_heads": 4, "resid_pdrop": 0.1, "embd_pdrop": 0.0, "attention_dropout": 0.0, "hidden_act": "gelu_new", "max_position_embeddings": 2048, "initializer_range": 0.02, "layer_norm_eps": 1e-05, "use_cache": true, "tie_word_embeddings": false, "rope_parameters": { "rope_theta": 10000.0, "partial_rotary_factor": 0.4, "rope_type": "default" }, "qk_layernorm": false, "bos_token_id": 50256, "eos_token_id": 50256, "pad_token_id": null, "model_type": "phi", "output_attentions": false, "dtype": "float16" }, "covers": { "microsoft/phi-2": "810d367871c1d460086d9f82db8696f2e0a0fcd0" }, "built_from": { "repo": "microsoft/phi-2", "revision": "810d367871c1d460086d9f82db8696f2e0a0fcd0" }, "seed": 17, "dtype": "float16", "builder": "example/build_fixtures.py", "builder_revision": "98ca7f9ba1ef19455052778369c6bd86c8c7689f", "recipe": { "seed": 17, "model_class": "PhiForCausalLM" }, "shrink": "rule", "model_class": "PhiForCausalLM", "parameters": 6671616, "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": "PhiForCausalLM" }, "weights": "random", "gate": { "random": { "pass": true, "accepted_insensitive": {}, "threshold": 0.001, "deltas": { "input_ids:roll_batch": 1.275390625, "input_ids:reverse": 1.21044921875, "attention_mask:all_ones": 0.837646484375 }, "weakest": "attention_mask:all_ones", "no_perturbation_defined": [], "not_a_forward_parameter": [] } }, "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": [ "CODE_OF_CONDUCT.md", "LICENSE", "NOTICE.md", "SECURITY.md", "added_tokens.json", "data_summary_card.md", "generation_config.json", "merges.txt", "special_tokens_map.json", "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": { "CODE_OF_CONDUCT.md": "9daeae709a0bd71bcfd1c96dc5822ecec5210327eff929da64b0ae7f8faf1444", "LICENSE": "fa8235e5b48faca34e3ca98cf4f694ef08bd216d28b58071a1f85b1d50cb814d", "NOTICE.md": "6ac541192c253d21cd16a597ce69c515868e031db3f32f75d996b993b5aca216", "SECURITY.md": "7b6976eec43edfa68b79a459dd089c56b7a395916dbf1a01bd11e6d86e12128f", "added_tokens.json": "eca8e90d1be6af3afeb1a37f070f9492645c551a9fb2e1ce2a4a8b21f04ffc8b", "config.json": "ff3e3c9cabbd0414e32f13e2a572432d2c0d5126009fab14054dd71b0220c348", "data_summary_card.md": "a12a5d701ee9e71bc1dde42e53c0e82842f38a6ecbebd595e3c72048c9fb499f", "generation_config.json": "ff2cfaa2f4c36c6ef17614096093e3b49c9efe6ddfd6cf20a50de5104ee2d1ed", "merges.txt": "1ce1664773c50f3e0cc8842619a93edc4624525b728b188a9e0be33b7726adc5", "model.safetensors": "001723778a24d6397593f66785c33144f93c495ab9a75220e340ad93f4c3fbc1", "special_tokens_map.json": "6f50ab5a5a509a1c309d6171f339b196a900dc9c99ad0408ff23bb615fdae7ad", "tokenizer.json": "337da36be7a71a6e88aa9148967a7bc8736f4b47c7de8e19ba92b89e80734cfc", "tokenizer_config.json": "1906c7514a738bd743a05d7d95c8e9aeddb4d8c9ffdda4fb6ee05653fb6b97c4", "vocab.json": "3ba3c3109ff33976c4bd966589c11ee14fcaa1f4c9e5e154c2ed7f99d80709e7" } }