tarekziade HF Staff
Publish MIC usage, slim fixture and reviewed Transformers expectations
1937d55 verified | { | |
| "fixture": "huggingface-mic/MiniMax-M2.5-mic", | |
| "key": "caec3890e5c190d78f66a8ca8f32c7ef4635c4e447de9d284911e5f9d304e46c", | |
| "key_fields": { | |
| "architectures": [ | |
| "MiniMaxM2ForCausalLM" | |
| ], | |
| "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": 200064, | |
| "hidden_size": 64, | |
| "intermediate_size": 128, | |
| "num_hidden_layers": 2, | |
| "num_attention_heads": 4, | |
| "num_key_value_heads": 2, | |
| "head_dim": 16, | |
| "hidden_act": "silu", | |
| "max_position_embeddings": 196608, | |
| "initializer_range": 0.02, | |
| "rms_norm_eps": 1e-06, | |
| "use_cache": true, | |
| "pad_token_id": null, | |
| "bos_token_id": 200034, | |
| "eos_token_id": 200020, | |
| "tie_word_embeddings": false, | |
| "attention_dropout": 0.0, | |
| "num_experts_per_tok": 2, | |
| "num_local_experts": 4, | |
| "output_router_logits": false, | |
| "router_aux_loss_coef": 0.001, | |
| "router_jitter_noise": 0.0, | |
| "rope_parameters": { | |
| "rope_theta": 5000000, | |
| "partial_rotary_factor": 0.5, | |
| "rope_type": "default" | |
| }, | |
| "attn_type_list": [ | |
| 1, | |
| 1 | |
| ], | |
| "auto_map": { | |
| "AutoConfig": "configuration_minimax_m2.MiniMaxM2Config", | |
| "AutoModelForCausalLM": "modeling_minimax_m2.MiniMaxM2ForCausalLM" | |
| }, | |
| "model_type": "minimax_m2", | |
| "mtp_transformer_layers": 1, | |
| "num_mtp_modules": 3, | |
| "qk_norm_type": "per_layer", | |
| "rotary_dim": 64, | |
| "scoring_func": "sigmoid", | |
| "shared_intermediate_size": 0, | |
| "use_mtp": true, | |
| "use_qk_norm": true, | |
| "use_routing_bias": true, | |
| "output_attentions": false | |
| }, | |
| "covers": { | |
| "MiniMaxAI/MiniMax-M2.5": "2515a3ea0afec1dd22d0c7794b3a91557804d195" | |
| }, | |
| "built_from": { | |
| "repo": "MiniMaxAI/MiniMax-M2.5", | |
| "revision": "2515a3ea0afec1dd22d0c7794b3a91557804d195" | |
| }, | |
| "seed": 17, | |
| "dtype": "float32", | |
| "builder": "example/build_fixtures.py", | |
| "builder_revision": "d35073c5ff9cb9829433654d300580c6830c0ce9", | |
| "recipe": { | |
| "seed": 17 | |
| }, | |
| "shrink": "rule", | |
| "model_class": "MiniMaxM2ForCausalLM", | |
| "parameters": 25830400, | |
| "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": "MiniMaxM2ForCausalLM" | |
| }, | |
| "weights": "random", | |
| "gate": { | |
| "random": { | |
| "pass": true, | |
| "accepted_insensitive": {}, | |
| "threshold": 0.001, | |
| "deltas": { | |
| "input_ids:roll_batch": 0.6353573203086853, | |
| "input_ids:reverse": 1.152644157409668, | |
| "attention_mask:all_ones": null | |
| }, | |
| "weakest": "input_ids:roll_batch", | |
| "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": [ | |
| "LICENSE-MODEL", | |
| "chat_template.jinja", | |
| "configuration_minimax_m2.py", | |
| "generation_config.json", | |
| "merges.txt", | |
| "modeling_minimax_m2.py", | |
| "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": { | |
| "LICENSE-MODEL": "278b29a14cf6b0cc60519617909af61fa4a6aca4ee928d6178cb9799785c5cc0", | |
| "chat_template.jinja": "f83cec52bb3a7e5a3cb4d6fd1dfdd6f4fe13b06c5aaa39a1115e70a417fc75f5", | |
| "config.json": "97f5771eb8343ef05bb1951653f7a9df30c2eda8b6bdfd787835578314f3ae88", | |
| "configuration_minimax_m2.py": "035abed680c1ead2d374621917f0bf6e0dc6647804dee0bf7c037d8fa74500c1", | |
| "generation_config.json": "2c24fe5507e260bb081727e2a14693d9a982942e721b28720c184d201bffb9dd", | |
| "merges.txt": "ca47ea60cf5bd48832586adc264f439d0ea4254176b24a95f85eaa9750e2b5f9", | |
| "model.safetensors": "8972b4200921f7faa54792b9ed07f41b68607627ce66b6c8db178ec0334cd1db", | |
| "modeling_minimax_m2.py": "99b170a2911b830a5b3f77262d5ea0d4a658d9ffb393cbe7c2b1bf87e2b0b818", | |
| "tokenizer.json": "757622126525aeeb131756849d93298070ff3f0319c455ec8c5bb0f6b1cebbe8", | |
| "tokenizer_config.json": "60e392fea8d47660584deb3c6a716aab7b9f11fed24bc494baf5eb7ba809e882", | |
| "vocab.json": "b44c066b5dc34c800c4e3ecbd85f3e95ce3bfdbf8a5fe30223e005175103578a" | |
| } | |
| } | |