tarekziade HF Staff
Publish MIC usage, slim fixture and reviewed Transformers expectations
af86835 verified | { | |
| "fixture": "huggingface-mic/DeepSeek-Coder-V2-Lite-Instruct-mic", | |
| "key": "465d8cbffd2f7c7aea9ba70c04586d73d9a68bffbfe76c8d0df39619f57e24e0", | |
| "key_fields": { | |
| "architectures": [ | |
| "DeepseekV2ForCausalLM" | |
| ], | |
| "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": 102400, | |
| "hidden_size": 64, | |
| "intermediate_size": 128, | |
| "num_hidden_layers": 2, | |
| "num_attention_heads": 4, | |
| "num_key_value_heads": 4, | |
| "hidden_act": "silu", | |
| "max_position_embeddings": 163840, | |
| "initializer_range": 0.02, | |
| "rms_norm_eps": 1e-06, | |
| "use_cache": true, | |
| "pad_token_id": null, | |
| "bos_token_id": 100000, | |
| "eos_token_id": 100001, | |
| "pretraining_tp": 1, | |
| "tie_word_embeddings": false, | |
| "rope_parameters": { | |
| "beta_fast": 32, | |
| "beta_slow": 1, | |
| "factor": 40, | |
| "mscale": 0.707, | |
| "mscale_all_dim": 0.707, | |
| "original_max_position_embeddings": 4096, | |
| "type": "yarn", | |
| "rope_theta": 10000, | |
| "rope_type": "yarn" | |
| }, | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "mlp_bias": false, | |
| "head_dim": 8, | |
| "first_k_dense_replace": 1, | |
| "kv_lora_rank": 16, | |
| "q_lora_rank": null, | |
| "n_group": 1, | |
| "n_routed_experts": 4, | |
| "output_router_logits": false, | |
| "n_shared_experts": 2, | |
| "qk_nope_head_dim": 16, | |
| "qk_rope_head_dim": 8, | |
| "routed_scaling_factor": 1.0, | |
| "topk_group": 1, | |
| "topk_method": "greedy", | |
| "norm_topk_prob": false, | |
| "v_head_dim": 16, | |
| "num_experts_per_tok": 2, | |
| "moe_intermediate_size": 32, | |
| "qk_head_dim": 192, | |
| "auto_map": { | |
| "AutoConfig": "configuration_deepseek.DeepseekV2Config", | |
| "AutoModel": "modeling_deepseek.DeepseekV2Model", | |
| "AutoModelForCausalLM": "modeling_deepseek.DeepseekV2ForCausalLM" | |
| }, | |
| "aux_loss_alpha": 0.001, | |
| "model_type": "deepseek_v2", | |
| "moe_layer_freq": 1, | |
| "scoring_func": "softmax", | |
| "seq_aux": true, | |
| "output_attentions": false, | |
| "dtype": "bfloat16" | |
| }, | |
| "covers": { | |
| "deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct": "e434a23f91ba5b4923cf6c9d9a238eb4a08e3a11" | |
| }, | |
| "built_from": { | |
| "repo": "deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct", | |
| "revision": "e434a23f91ba5b4923cf6c9d9a238eb4a08e3a11" | |
| }, | |
| "seed": 17, | |
| "dtype": "bfloat16", | |
| "builder": "example/build_fixtures.py", | |
| "builder_revision": "d6c4e1368e76b3cb3fbd52f1d7813a1c321195c5", | |
| "recipe": { | |
| "seed": 17, | |
| "model_class": "DeepseekV2ForCausalLM" | |
| }, | |
| "shrink": "rule", | |
| "model_class": "DeepseekV2ForCausalLM", | |
| "parameters": 13196896, | |
| "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": "DeepseekV2ForCausalLM" | |
| }, | |
| "weights": "random", | |
| "gate": { | |
| "random": { | |
| "pass": true, | |
| "accepted_insensitive": {}, | |
| "threshold": 0.001, | |
| "deltas": { | |
| "input_ids:roll_batch": 1.11328125, | |
| "input_ids:reverse": 1.1640625, | |
| "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": [ | |
| "configuration_deepseek.py", | |
| "generation_config.json", | |
| "modeling_deepseek.py", | |
| "tokenization_deepseek_fast.py", | |
| "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": "130ea0a870be68586a9751f12e29a349d85ccbe9b38e8c109774a5e310a3d616", | |
| "configuration_deepseek.py": "6de5e7445ab490a16482027b131d5bd6a003da9b58a182a11432115fac84ce59", | |
| "generation_config.json": "63ec07f6cb36bd8359b44f671123518dd6fbf55007e0c74aa96c296a75def8b5", | |
| "model.safetensors": "ac89c50e9c9eb2d80589a878e2a3ba86f089c8a9005e87e8a7271034e8fd15c6", | |
| "modeling_deepseek.py": "7d8e5221095286eea991137760893fd7ba52727c0b4ebf48ec09e8bc56b45b9c", | |
| "tokenization_deepseek_fast.py": "702fc7dbadc450f2fecfbc59e248ecca5845b49101e8f58f805261c920a3effa", | |
| "tokenizer.json": "091b9dadb9845f0e8386c38bdb87e98db8adc7b0aacf36cb9257c00d0a668714", | |
| "tokenizer_config.json": "31181eaf79394ea26728d95ecb54fe7c8413e6f56085dbabc8b0818134380ec8" | |
| } | |
| } | |