{ "_name_or_path": "distributed/llama-1b", "all_reduce_scores": { "0": "NOT_ALIVE", "1": "NOT_ALIVE", "10": "NON_PARTICIPATING", "100": "NOT_ALIVE", "101": "NOT_ALIVE", "102": "FAIL", "103": "NOT_ALIVE", "104": "NOT_ALIVE", "105": "FAIL", "106": "FAIL", "107": "NOT_ALIVE", "108": "NOT_ALIVE", "109": "NOT_ALIVE", "11": "NON_PARTICIPATING", "110": "NON_PARTICIPATING", "111": "NON_PARTICIPATING", "112": "NON_PARTICIPATING", "113": "FAIL", "114": "NOT_ALIVE", "115": "FAIL", "116": "NON_PARTICIPATING", "117": "NON_PARTICIPATING", "118": "NON_PARTICIPATING", "119": "FAIL", "12": "NOT_ALIVE", "120": "NON_PARTICIPATING", "121": "NOT_ALIVE", "122": "NON_PARTICIPATING", "123": "FAIL", "124": "FAIL", "125": "NON_PARTICIPATING", "126": "NON_PARTICIPATING", "127": "FAIL", "128": "NON_PARTICIPATING", "129": "NOT_ALIVE", "13": "FAIL", "130": "NON_PARTICIPATING", "131": "NOT_ALIVE", "132": "NOT_ALIVE", "133": "SUCCESS", "134": "NON_PARTICIPATING", "135": "FAIL", "136": "NOT_ALIVE", "137": "NOT_ALIVE", "138": "NON_PARTICIPATING", "139": "NOT_ALIVE", "14": "FAIL", "140": "NOT_ALIVE", "141": "NOT_ALIVE", "142": "NON_PARTICIPATING", "143": "NOT_ALIVE", "144": "NON_PARTICIPATING", "145": "NOT_ALIVE", "146": "NON_PARTICIPATING", "147": "FAIL", "148": "FAIL", "149": "NON_PARTICIPATING", "15": "NOT_ALIVE", "150": "NOT_ALIVE", "151": "NOT_ALIVE", "152": "NOT_ALIVE", "153": "NON_PARTICIPATING", "154": "NOT_ALIVE", "155": "NON_PARTICIPATING", "156": "NON_PARTICIPATING", "157": "NON_PARTICIPATING", "158": "NON_PARTICIPATING", "159": "NON_PARTICIPATING", "16": "NON_PARTICIPATING", "160": "NOT_ALIVE", "161": "NOT_ALIVE", "162": "NON_PARTICIPATING", "163": "NON_PARTICIPATING", "164": "NOT_ALIVE", "165": "NON_PARTICIPATING", "166": "NOT_ALIVE", "167": "NON_PARTICIPATING", "168": "NON_PARTICIPATING", "169": "NON_PARTICIPATING", "17": "NON_PARTICIPATING", "170": "FAIL", "171": "NOT_ALIVE", "172": "NOT_ALIVE", "173": "NOT_ALIVE", "174": "NON_PARTICIPATING", "175": "NON_PARTICIPATING", "176": "NOT_ALIVE", "177": "FAIL", "178": "NOT_ALIVE", "179": "NOT_ALIVE", "18": "NON_PARTICIPATING", "180": "FAIL", "181": "NOT_ALIVE", "182": "NON_PARTICIPATING", "183": "NOT_ALIVE", "184": "FAIL", "185": "FAIL", "186": "SUCCESS", "187": "NOT_ALIVE", "188": "FAIL", "189": "NOT_ALIVE", "19": "NON_PARTICIPATING", "190": "FAIL", "191": "NON_PARTICIPATING", "192": "NON_PARTICIPATING", "193": "NON_PARTICIPATING", "194": "NOT_ALIVE", "195": "NON_PARTICIPATING", "196": "FAIL", "197": "NON_PARTICIPATING", "198": "NON_PARTICIPATING", "199": "FAIL", "2": "FAIL", "20": "NON_PARTICIPATING", "200": "FAIL", "201": "NON_PARTICIPATING", "202": "NOT_ALIVE", "203": "NOT_ALIVE", "204": "NOT_ALIVE", "205": "NOT_ALIVE", "206": "FAIL", "207": "NOT_ALIVE", "208": "NON_PARTICIPATING", "209": "FAIL", "21": "FAIL", "210": "NON_PARTICIPATING", "211": "NON_PARTICIPATING", "212": "FAIL", "213": "NON_PARTICIPATING", "214": "NOT_ALIVE", "215": "NOT_ALIVE", "216": "NOT_ALIVE", "217": "NOT_ALIVE", "218": "NOT_ALIVE", "219": "SUCCESS", "22": "NOT_ALIVE", "220": "NOT_ALIVE", "221": "NON_PARTICIPATING", "222": "NOT_ALIVE", "223": "NON_PARTICIPATING", "224": "NOT_ALIVE", "225": "NON_PARTICIPATING", "226": "FAIL", "227": "FAIL", "228": "NOT_ALIVE", "229": "NON_PARTICIPATING", "23": "NOT_ALIVE", "230": "NON_PARTICIPATING", "231": "NOT_ALIVE", "232": "NOT_ALIVE", "233": "NON_PARTICIPATING", "234": "NON_PARTICIPATING", "235": "NON_PARTICIPATING", "236": "NON_PARTICIPATING", "237": "NON_PARTICIPATING", "238": "NON_PARTICIPATING", "239": "NON_PARTICIPATING", "24": "NON_PARTICIPATING", "240": "NOT_ALIVE", "241": "NON_PARTICIPATING", "242": "NON_PARTICIPATING", "243": "NOT_ALIVE", "244": "NOT_ALIVE", "245": "FAIL", "246": "NON_PARTICIPATING", "247": "NON_PARTICIPATING", "248": "FAIL", "249": "NOT_ALIVE", "25": "SUCCESS", "250": "FAIL", "251": "NOT_ALIVE", "252": "FAIL", "253": "FAIL", "254": "NOT_ALIVE", "255": "NON_PARTICIPATING", "26": "FAIL", "27": "NOT_ALIVE", "28": "NON_PARTICIPATING", "29": "NON_PARTICIPATING", "3": "NON_PARTICIPATING", "30": "NON_PARTICIPATING", "31": "NOT_ALIVE", "32": "NON_PARTICIPATING", "33": "NOT_ALIVE", "34": "NON_PARTICIPATING", "35": "FAIL", "36": "SUCCESS", "37": "NOT_ALIVE", "38": "NOT_ALIVE", "39": "NON_PARTICIPATING", "4": "SUCCESS", "40": "NON_PARTICIPATING", "41": "FAIL", "42": "SUCCESS", "43": "NOT_ALIVE", "44": "FAIL", "45": "NOT_ALIVE", "46": "NOT_ALIVE", "47": "NON_PARTICIPATING", "48": "NON_PARTICIPATING", "49": "NOT_ALIVE", "5": "NON_PARTICIPATING", "50": "NOT_ALIVE", "51": "FAIL", "52": "NOT_ALIVE", "53": "FAIL", "54": "NOT_ALIVE", "55": "NOT_ALIVE", "56": "NOT_ALIVE", "57": "FAIL", "58": "NOT_ALIVE", "59": "NOT_ALIVE", "6": "NOT_ALIVE", "60": "NOT_ALIVE", "61": "NON_PARTICIPATING", "62": "NON_PARTICIPATING", "63": "NON_PARTICIPATING", "64": "FAIL", "65": "NOT_ALIVE", "66": "SUCCESS", "67": "NON_PARTICIPATING", "68": "NOT_ALIVE", "69": "FAIL", "7": "NOT_ALIVE", "70": "NOT_ALIVE", "71": "NOT_ALIVE", "72": "NON_PARTICIPATING", "73": "NON_PARTICIPATING", "74": "FAIL", "75": "NON_PARTICIPATING", "76": "NON_PARTICIPATING", "77": "NOT_ALIVE", "78": "NON_PARTICIPATING", "79": "NON_PARTICIPATING", "8": "NOT_ALIVE", "80": "NOT_ALIVE", "81": "NOT_ALIVE", "82": "NOT_ALIVE", "83": "NON_PARTICIPATING", "84": "FAIL", "85": "NON_PARTICIPATING", "86": "NOT_ALIVE", "87": "NON_PARTICIPATING", "88": "NON_PARTICIPATING", "89": "FAIL", "9": "SUCCESS", "90": "NON_PARTICIPATING", "91": "NON_PARTICIPATING", "92": "SUCCESS", "93": "NOT_ALIVE", "94": "FAIL", "95": "NOT_ALIVE", "96": "NON_PARTICIPATING", "97": "NON_PARTICIPATING", "98": "NON_PARTICIPATING", "99": "NON_PARTICIPATING" }, "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "block_list": [ 6468728, 6468758 ], "bos_token_id": 1, "eos_token_id": 2, "hidden_act": "silu", "hidden_size": 2048, "initializer_range": 0.02, "inner_step": 59, "intermediate_size": 5632, "last_allreduce_block": 6468297, "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 }