{ "_name_or_path": "distributed/llama-1b", "all_reduce_scores": { "0": "NOT_ALIVE", "1": "SUCCESS", "10": "NOT_ALIVE", "100": "NOT_ALIVE", "101": "NOT_ALIVE", "102": "NON_PARTICIPATING", "103": "NON_PARTICIPATING", "104": "NOT_ALIVE", "105": "NON_PARTICIPATING", "106": "NOT_ALIVE", "107": "NOT_ALIVE", "108": "NON_PARTICIPATING", "109": "NOT_ALIVE", "11": "NOT_ALIVE", "110": "NOT_ALIVE", "111": "NOT_ALIVE", "112": "NOT_ALIVE", "113": "NOT_ALIVE", "114": "NOT_ALIVE", "115": "NOT_ALIVE", "116": "NOT_ALIVE", "117": "NOT_ALIVE", "118": "NON_PARTICIPATING", "119": "NOT_ALIVE", "12": "NOT_ALIVE", "120": "NOT_ALIVE", "121": "NOT_ALIVE", "122": "NON_PARTICIPATING", "123": "NOT_ALIVE", "124": "NON_PARTICIPATING", "125": "NOT_ALIVE", "126": "NOT_ALIVE", "127": "NOT_ALIVE", "128": "NOT_ALIVE", "129": "NOT_ALIVE", "13": "NOT_ALIVE", "130": "NOT_ALIVE", "131": "NON_PARTICIPATING", "132": "NOT_ALIVE", "133": "NOT_ALIVE", "134": "NON_PARTICIPATING", "135": "NON_PARTICIPATING", "136": "NON_PARTICIPATING", "137": "NON_PARTICIPATING", "138": "NON_PARTICIPATING", "139": "NON_PARTICIPATING", "14": "NOT_ALIVE", "140": "NON_PARTICIPATING", "141": "NON_PARTICIPATING", "142": "NON_PARTICIPATING", "143": "NON_PARTICIPATING", "144": "SUCCESS", "145": "NOT_ALIVE", "146": "NON_PARTICIPATING", "147": "NON_PARTICIPATING", "148": "NON_PARTICIPATING", "149": "NOT_ALIVE", "15": "SUCCESS", "150": "NOT_ALIVE", "151": "NON_PARTICIPATING", "152": "NOT_ALIVE", "153": "NOT_ALIVE", "154": "NON_PARTICIPATING", "155": "NON_PARTICIPATING", "156": "NOT_ALIVE", "157": "NOT_ALIVE", "158": "NON_PARTICIPATING", "159": "NON_PARTICIPATING", "16": "NOT_ALIVE", "160": "NON_PARTICIPATING", "161": "NON_PARTICIPATING", "162": "NON_PARTICIPATING", "163": "NON_PARTICIPATING", "164": "NOT_ALIVE", "165": "NON_PARTICIPATING", "166": "NON_PARTICIPATING", "167": "NOT_ALIVE", "168": "NON_PARTICIPATING", "169": "NON_PARTICIPATING", "17": "NOT_ALIVE", "170": "NON_PARTICIPATING", "171": "NON_PARTICIPATING", "172": "NON_PARTICIPATING", "173": "NON_PARTICIPATING", "174": "NON_PARTICIPATING", "175": "NON_PARTICIPATING", "176": "NOT_ALIVE", "177": "NOT_ALIVE", "178": "NOT_ALIVE", "179": "NON_PARTICIPATING", "18": "SUCCESS", "180": "NON_PARTICIPATING", "181": "NOT_ALIVE", "182": "NON_PARTICIPATING", "183": "NOT_ALIVE", "184": "NON_PARTICIPATING", "185": "NON_PARTICIPATING", "186": "NOT_ALIVE", "187": "NON_PARTICIPATING", "188": "SUCCESS", "189": "SUCCESS", "19": "NOT_ALIVE", "190": "NON_PARTICIPATING", "191": "NOT_ALIVE", "192": "NON_PARTICIPATING", "193": "NOT_ALIVE", "194": "NOT_ALIVE", "195": "NOT_ALIVE", "196": "FAIL", "197": "NOT_ALIVE", "198": "NON_PARTICIPATING", "199": "NOT_ALIVE", "2": "NOT_ALIVE", "20": "NOT_ALIVE", "200": "NON_PARTICIPATING", "201": "SUCCESS", "202": "NON_PARTICIPATING", "203": "NOT_ALIVE", "204": "NOT_ALIVE", "205": "NON_PARTICIPATING", "206": "NOT_ALIVE", "207": "NOT_ALIVE", "208": "NON_PARTICIPATING", "209": "FAIL", "21": "NOT_ALIVE", "210": "NOT_ALIVE", "211": "NON_PARTICIPATING", "212": "NON_PARTICIPATING", "213": "NON_PARTICIPATING", "214": "NON_PARTICIPATING", "215": "NON_PARTICIPATING", "216": "NON_PARTICIPATING", "217": "SUCCESS", "218": "NOT_ALIVE", "219": "NOT_ALIVE", "22": "NOT_ALIVE", "220": "NOT_ALIVE", "221": "NON_PARTICIPATING", "222": "NOT_ALIVE", "223": "NOT_ALIVE", "224": "NOT_ALIVE", "225": "NON_PARTICIPATING", "226": "NON_PARTICIPATING", "227": "NON_PARTICIPATING", "228": "SUCCESS", "229": "NOT_ALIVE", "23": "FAIL", "230": "SUCCESS", "231": "NOT_ALIVE", "232": "NOT_ALIVE", "233": "NON_PARTICIPATING", "234": "NON_PARTICIPATING", "235": "NON_PARTICIPATING", "236": "NOT_ALIVE", "237": "NON_PARTICIPATING", "238": "NOT_ALIVE", "239": "NON_PARTICIPATING", "24": "NOT_ALIVE", "240": "NOT_ALIVE", "241": "NON_PARTICIPATING", "242": "FAIL", "243": "NOT_ALIVE", "244": "SUCCESS", "245": "NON_PARTICIPATING", "246": "NON_PARTICIPATING", "247": "SUCCESS", "248": "NOT_ALIVE", "249": "NOT_ALIVE", "25": "SUCCESS", "250": "NOT_ALIVE", "251": "NOT_ALIVE", "252": "NON_PARTICIPATING", "253": "NON_PARTICIPATING", "254": "NON_PARTICIPATING", "255": "NON_PARTICIPATING", "26": "NON_PARTICIPATING", "27": "NOT_ALIVE", "28": "NON_PARTICIPATING", "29": "NOT_ALIVE", "3": "SUCCESS", "30": "NOT_ALIVE", "31": "SUCCESS", "32": "SUCCESS", "33": "NOT_ALIVE", "34": "NON_PARTICIPATING", "35": "NON_PARTICIPATING", "36": "NON_PARTICIPATING", "37": "NOT_ALIVE", "38": "NON_PARTICIPATING", "39": "SUCCESS", "4": "NOT_ALIVE", "40": "NON_PARTICIPATING", "41": "NOT_ALIVE", "42": "NOT_ALIVE", "43": "NON_PARTICIPATING", "44": "NON_PARTICIPATING", "45": "NOT_ALIVE", "46": "NON_PARTICIPATING", "47": "NOT_ALIVE", "48": "NON_PARTICIPATING", "49": "NOT_ALIVE", "5": "NOT_ALIVE", "50": "NOT_ALIVE", "51": "NON_PARTICIPATING", "52": "NOT_ALIVE", "53": "NON_PARTICIPATING", "54": "NOT_ALIVE", "55": "NON_PARTICIPATING", "56": "NON_PARTICIPATING", "57": "NON_PARTICIPATING", "58": "NOT_ALIVE", "59": "NON_PARTICIPATING", "6": "NOT_ALIVE", "60": "NON_PARTICIPATING", "61": "NON_PARTICIPATING", "62": "NOT_ALIVE", "63": "NON_PARTICIPATING", "64": "NON_PARTICIPATING", "65": "NON_PARTICIPATING", "66": "NON_PARTICIPATING", "67": "NON_PARTICIPATING", "68": "NOT_ALIVE", "69": "NOT_ALIVE", "7": "NOT_ALIVE", "70": "NON_PARTICIPATING", "71": "NON_PARTICIPATING", "72": "NOT_ALIVE", "73": "NOT_ALIVE", "74": "NOT_ALIVE", "75": "NON_PARTICIPATING", "76": "NOT_ALIVE", "77": "NON_PARTICIPATING", "78": "SUCCESS", "79": "NOT_ALIVE", "8": "NOT_ALIVE", "80": "NOT_ALIVE", "81": "NON_PARTICIPATING", "82": "NOT_ALIVE", "83": "SUCCESS", "84": "NON_PARTICIPATING", "85": "NOT_ALIVE", "86": "NON_PARTICIPATING", "87": "NOT_ALIVE", "88": "NON_PARTICIPATING", "89": "NOT_ALIVE", "9": "NOT_ALIVE", "90": "NOT_ALIVE", "91": "NOT_ALIVE", "92": "NOT_ALIVE", "93": "NON_PARTICIPATING", "94": "NON_PARTICIPATING", "95": "NON_PARTICIPATING", "96": "NON_PARTICIPATING", "97": "NON_PARTICIPATING", "98": "NON_PARTICIPATING", "99": "NON_PARTICIPATING" }, "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "block_list": [ 6127652, 6127668 ], "bos_token_id": 1, "eos_token_id": 2, "hidden_act": "silu", "hidden_size": 2048, "initializer_range": 0.02, "inner_step": 49, "intermediate_size": 5632, "last_allreduce_block": 6127267, "max_position_embeddings": 2048, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 22, "num_key_value_heads": 4, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 10000.0, "tie_word_embeddings": false, "torch_dtype": "float32", "transformers_version": "4.39.3", "use_cache": false, "vocab_size": 32000 }