{ "fixture": "huggingface-mic/tiny-random-aya-base-mic", "key": "942d7e0ff56152ffba0f6968c631c68ce6925551b97069225beed6880b8f4b5d", "key_fields": { "architectures": [ "Cohere2ForCausalLM" ], "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": 262144, "hidden_size": 64, "intermediate_size": 128, "logit_scale": 1.0, "num_hidden_layers": 2, "num_attention_heads": 4, "num_key_value_heads": 4, "hidden_act": "silu", "max_position_embeddings": 500000, "initializer_range": 0.02, "layer_norm_eps": 1e-05, "use_cache": true, "pad_token_id": 0, "bos_token_id": 2, "eos_token_id": 3, "tie_word_embeddings": true, "rope_parameters": { "rope_theta": 50000, "rope_type": "default" }, "attention_bias": false, "attention_dropout": 0.0, "sliding_window": 4096, "layer_types": [ "sliding_attention", "full_attention" ], "head_dim": 16, "_sliding_window_pattern": 4, "cache_implementation": "hybrid", "layer_switch": 4, "model_type": "cohere2", "order_of_interleaved_layers": "local_attn_first", "position_embedding_type": "rope_gptj", "rotary_pct": 1.0, "sliding_window_pattern": 4, "use_embedding_sharing": true, "use_gated_activation": true, "use_parallel_block": true, "use_parallel_embedding": false, "use_qk_norm": false, "output_attentions": false }, "covers": { "optimum-intel-internal-testing/tiny-random-aya-base": "e85478d37f912454b23ff2a9c7af16a960837644" }, "built_from": { "repo": "optimum-intel-internal-testing/tiny-random-aya-base", "revision": "e85478d37f912454b23ff2a9c7af16a960837644" }, "seed": 17, "dtype": "float32", "builder": "example/build_fixtures.py", "builder_revision": "d1b2e0637b293ce867d0364161d6d69777d0941b", "recipe": { "seed": 17, "model_class": "Cohere2ForCausalLM" }, "shrink": "rule", "model_class": "Cohere2ForCausalLM", "parameters": 16859328, "missing_parameters": [], "serialization": { "file": "model.safetensors", "reverted_conversions": [] }, "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": "Cohere2ForCausalLM" }, "weights": "random", "gate": { "random": { "pass": true, "accepted_insensitive": {}, "threshold": 0.001, "deltas": { "input_ids:roll_batch": 1.3876315355300903, "input_ids:reverse": 1.5412429571151733, "attention_mask:all_ones": 0.4773968458175659 }, "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": [ "generation_config.json", "special_tokens_map.json", "tokenizer.json", "tokenizer_config.json" ], "environment": { "python": "3.12.12", "machine": "arm64", "torch": "2.14.1", "transformers": "5.22.0.dev0", "tokenizers": "0.23.3" }, "sha256": { "config.json": "c138ff718afd9e1a5a19804e88225da79e5397bdf0fa5bef136207f189105349", "generation_config.json": "fa1f6af5005a20eafe7ce1a4365134a2e7a791d739e501e981946bd71f417d64", "model.safetensors": "c4cf5b8ea901291cfeda11936929fa25869fc600d2529620e9307241c6cc6a06", "special_tokens_map.json": "2b7ce39630f8a18705331a9b32c3b2236e96e7410791f98a9f8b6494e3ab2ea6", "tokenizer.json": "8f21f6c4f761c192f486ea2c5b06b62b3ef30819b33dc105bdf8b26c8e7974f6", "tokenizer_config.json": "a6f3636edc7f2b5db4fa243a9b1d540ff6fc651f6da4fc6afed6fc2332451bd7" } }