{ "_name_or_path": "distributed/llama-1b", "all_reduce_scores": { "0": "NOT_ALIVE", "1": "SUCCESS", "10": "NOT_ALIVE", "100": "NOT_ALIVE", "101": "NON_PARTICIPATING", "102": "NOT_ALIVE", "103": "NON_PARTICIPATING", "104": "NOT_ALIVE", "105": "NON_PARTICIPATING", "106": "NOT_ALIVE", "107": "NON_PARTICIPATING", "108": "NON_PARTICIPATING", "109": "SUCCESS", "11": "NOT_ALIVE", "110": "NOT_ALIVE", "111": "SUCCESS", "112": "NOT_ALIVE", "113": "SUCCESS", "114": "NON_PARTICIPATING", "115": "NON_PARTICIPATING", "116": "NON_PARTICIPATING", "117": "SUCCESS", "118": "NOT_ALIVE", "119": "NON_PARTICIPATING", "12": "NOT_ALIVE", "120": "NOT_ALIVE", "121": "NON_PARTICIPATING", "122": "NON_PARTICIPATING", "123": "NON_PARTICIPATING", "124": "NON_PARTICIPATING", "125": "NOT_ALIVE", "126": "NOT_ALIVE", "127": "SUCCESS", "128": "NON_PARTICIPATING", "129": "NON_PARTICIPATING", "13": "SUCCESS", "130": "NON_PARTICIPATING", "131": "NON_PARTICIPATING", "132": "NOT_ALIVE", "133": "SUCCESS", "134": "NON_PARTICIPATING", "135": "NOT_ALIVE", "136": "NON_PARTICIPATING", "137": "NON_PARTICIPATING", "138": "NON_PARTICIPATING", "139": "SUCCESS", "14": "NOT_ALIVE", "140": "NOT_ALIVE", "141": "NOT_ALIVE", "142": "NON_PARTICIPATING", "143": "NON_PARTICIPATING", "144": "NON_PARTICIPATING", "145": "NOT_ALIVE", "146": "NON_PARTICIPATING", "147": "NON_PARTICIPATING", "148": "NON_PARTICIPATING", "149": "NON_PARTICIPATING", "15": "NOT_ALIVE", "150": "NON_PARTICIPATING", "151": "NON_PARTICIPATING", "152": "SUCCESS", "153": "NON_PARTICIPATING", "154": "NOT_ALIVE", "155": "NON_PARTICIPATING", "156": "NOT_ALIVE", "157": "NOT_ALIVE", "158": "SUCCESS", "159": "NOT_ALIVE", "16": "FAIL", "160": "SUCCESS", "161": "NOT_ALIVE", "162": "NON_PARTICIPATING", "163": "NON_PARTICIPATING", "164": "NON_PARTICIPATING", "165": "SUCCESS", "166": "NOT_ALIVE", "167": "NON_PARTICIPATING", "168": "SUCCESS", "169": "NOT_ALIVE", "17": "NOT_ALIVE", "170": "NON_PARTICIPATING", "171": "SUCCESS", "172": "NOT_ALIVE", "173": "SUCCESS", "174": "NON_PARTICIPATING", "175": "NOT_ALIVE", "176": "NOT_ALIVE", "177": "NON_PARTICIPATING", "178": "NOT_ALIVE", "179": "NON_PARTICIPATING", "18": "NOT_ALIVE", "180": "NOT_ALIVE", "181": "NON_PARTICIPATING", "182": "SUCCESS", "183": "SUCCESS", "184": "SUCCESS", "185": "SUCCESS", "186": "NOT_ALIVE", "187": "NON_PARTICIPATING", "188": "NON_PARTICIPATING", "189": "NON_PARTICIPATING", "19": "NOT_ALIVE", "190": "SUCCESS", "191": "NOT_ALIVE", "192": "NON_PARTICIPATING", "193": "NON_PARTICIPATING", "194": "SUCCESS", "195": "NON_PARTICIPATING", "196": "SUCCESS", "197": "NOT_ALIVE", "198": "NON_PARTICIPATING", "199": "SUCCESS", "2": "NOT_ALIVE", "20": "NOT_ALIVE", "200": "SUCCESS", "201": "NOT_ALIVE", "202": "NON_PARTICIPATING", "203": "FAIL", "204": "NOT_ALIVE", "205": "NON_PARTICIPATING", "206": "NOT_ALIVE", "207": "NON_PARTICIPATING", "208": "NON_PARTICIPATING", "209": "SUCCESS", "21": "NOT_ALIVE", "210": "FAIL", "211": "SUCCESS", "212": "NOT_ALIVE", "213": "NON_PARTICIPATING", "214": "SUCCESS", "215": "NON_PARTICIPATING", "216": "SUCCESS", "217": "SUCCESS", "218": "NON_PARTICIPATING", "219": "NON_PARTICIPATING", "22": "NOT_ALIVE", "220": "SUCCESS", "221": "NON_PARTICIPATING", "222": "NON_PARTICIPATING", "223": "NON_PARTICIPATING", "224": "NON_PARTICIPATING", "225": "NON_PARTICIPATING", "226": "NON_PARTICIPATING", "227": "NOT_ALIVE", "228": "SUCCESS", "229": "NOT_ALIVE", "23": "NOT_ALIVE", "230": "NON_PARTICIPATING", "231": "SUCCESS", "232": "NOT_ALIVE", "233": "NOT_ALIVE", "234": "SUCCESS", "235": "NON_PARTICIPATING", "236": "NON_PARTICIPATING", "237": "SUCCESS", "238": "NOT_ALIVE", "239": "NON_PARTICIPATING", "24": "NOT_ALIVE", "240": "NON_PARTICIPATING", "241": "NON_PARTICIPATING", "242": "NON_PARTICIPATING", "243": "NOT_ALIVE", "244": "NON_PARTICIPATING", "245": "NON_PARTICIPATING", "246": "NON_PARTICIPATING", "247": "SUCCESS", "248": "SUCCESS", "249": "SUCCESS", "25": "SUCCESS", "250": "NOT_ALIVE", "251": "SUCCESS", "252": "NOT_ALIVE", "253": "NON_PARTICIPATING", "254": "NOT_ALIVE", "255": "NON_PARTICIPATING", "26": "NON_PARTICIPATING", "27": "SUCCESS", "28": "NON_PARTICIPATING", "29": "NON_PARTICIPATING", "3": "NOT_ALIVE", "30": "NOT_ALIVE", "31": "NOT_ALIVE", "32": "NON_PARTICIPATING", "33": "NON_PARTICIPATING", "34": "NOT_ALIVE", "35": "NON_PARTICIPATING", "36": "SUCCESS", "37": "NOT_ALIVE", "38": "NON_PARTICIPATING", "39": "NON_PARTICIPATING", "4": "SUCCESS", "40": "NOT_ALIVE", "41": "NON_PARTICIPATING", "42": "SUCCESS", "43": "SUCCESS", "44": "NON_PARTICIPATING", "45": "NOT_ALIVE", "46": "SUCCESS", "47": "NON_PARTICIPATING", "48": "NON_PARTICIPATING", "49": "NOT_ALIVE", "5": "NOT_ALIVE", "50": "SUCCESS", "51": "NON_PARTICIPATING", "52": "NON_PARTICIPATING", "53": "NON_PARTICIPATING", "54": "NON_PARTICIPATING", "55": "NON_PARTICIPATING", "56": "NON_PARTICIPATING", "57": "NON_PARTICIPATING", "58": "NOT_ALIVE", "59": "NOT_ALIVE", "6": "NOT_ALIVE", "60": "SUCCESS", "61": "NON_PARTICIPATING", "62": "NOT_ALIVE", "63": "NOT_ALIVE", "64": "NON_PARTICIPATING", "65": "SUCCESS", "66": "NON_PARTICIPATING", "67": "NON_PARTICIPATING", "68": "NON_PARTICIPATING", "69": "SUCCESS", "7": "NOT_ALIVE", "70": "NOT_ALIVE", "71": "SUCCESS", "72": "FAIL", "73": "NON_PARTICIPATING", "74": "SUCCESS", "75": "NON_PARTICIPATING", "76": "NON_PARTICIPATING", "77": "NON_PARTICIPATING", "78": "NON_PARTICIPATING", "79": "NOT_ALIVE", "8": "NOT_ALIVE", "80": "NOT_ALIVE", "81": "NON_PARTICIPATING", "82": "SUCCESS", "83": "NON_PARTICIPATING", "84": "NOT_ALIVE", "85": "NON_PARTICIPATING", "86": "NOT_ALIVE", "87": "NON_PARTICIPATING", "88": "NON_PARTICIPATING", "89": "NOT_ALIVE", "9": "NOT_ALIVE", "90": "NON_PARTICIPATING", "91": "SUCCESS", "92": "NOT_ALIVE", "93": "NOT_ALIVE", "94": "SUCCESS", "95": "NON_PARTICIPATING", "96": "SUCCESS", "97": "NON_PARTICIPATING", "98": "NOT_ALIVE", "99": "SUCCESS" }, "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "block_list": [ 6437377, 6437385 ], "bos_token_id": 1, "eos_token_id": 2, "hidden_act": "silu", "hidden_size": 2048, "initializer_range": 0.02, "inner_step": 292, "intermediate_size": 5632, "last_allreduce_block": 6435370, "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 }