{ "_name_or_path": "distributed/llama-1b", "all_reduce_scores": { "0": "NOT_ALIVE", "1": "NON_PARTICIPATING", "10": "NOT_ALIVE", "100": "NOT_ALIVE", "101": "NON_PARTICIPATING", "102": "NON_PARTICIPATING", "103": "NOT_ALIVE", "104": "NON_PARTICIPATING", "105": "NON_PARTICIPATING", "106": "NON_PARTICIPATING", "107": "NOT_ALIVE", "108": "NOT_ALIVE", "109": "NOT_ALIVE", "11": "NON_PARTICIPATING", "110": "NON_PARTICIPATING", "111": "NON_PARTICIPATING", "112": "SUCCESS", "113": "NOT_ALIVE", "114": "NON_PARTICIPATING", "115": "NOT_ALIVE", "116": "NON_PARTICIPATING", "117": "NOT_ALIVE", "118": "NOT_ALIVE", "119": "NON_PARTICIPATING", "12": "NON_PARTICIPATING", "120": "NON_PARTICIPATING", "121": "NOT_ALIVE", "122": "NON_PARTICIPATING", "123": "NOT_ALIVE", "124": "NON_PARTICIPATING", "125": "NON_PARTICIPATING", "126": "NOT_ALIVE", "127": "NOT_ALIVE", "128": "NOT_ALIVE", "129": "NON_PARTICIPATING", "13": "NOT_ALIVE", "130": "NON_PARTICIPATING", "131": "NOT_ALIVE", "132": "NON_PARTICIPATING", "133": "SUCCESS", "134": "NON_PARTICIPATING", "135": "NON_PARTICIPATING", "136": "SUCCESS", "137": "NOT_ALIVE", "138": "NON_PARTICIPATING", "139": "NON_PARTICIPATING", "14": "SUCCESS", "140": "NOT_ALIVE", "141": "NOT_ALIVE", "142": "NON_PARTICIPATING", "143": "NON_PARTICIPATING", "144": "NON_PARTICIPATING", "145": "NON_PARTICIPATING", "146": "NON_PARTICIPATING", "147": "NON_PARTICIPATING", "148": "NON_PARTICIPATING", "149": "NON_PARTICIPATING", "15": "NOT_ALIVE", "150": "NON_PARTICIPATING", "151": "NOT_ALIVE", "152": "NOT_ALIVE", "153": "SUCCESS", "154": "SUCCESS", "155": "SUCCESS", "156": "NOT_ALIVE", "157": "NOT_ALIVE", "158": "NON_PARTICIPATING", "159": "SUCCESS", "16": "NOT_ALIVE", "160": "NOT_ALIVE", "161": "NOT_ALIVE", "162": "NON_PARTICIPATING", "163": "NOT_ALIVE", "164": "SUCCESS", "165": "SUCCESS", "166": "NON_PARTICIPATING", "167": "NOT_ALIVE", "168": "NON_PARTICIPATING", "169": "NOT_ALIVE", "17": "NOT_ALIVE", "170": "NON_PARTICIPATING", "171": "NOT_ALIVE", "172": "NOT_ALIVE", "173": "SUCCESS", "174": "NOT_ALIVE", "175": "NOT_ALIVE", "176": "NOT_ALIVE", "177": "NOT_ALIVE", "178": "NON_PARTICIPATING", "179": "NOT_ALIVE", "18": "NOT_ALIVE", "180": "NOT_ALIVE", "181": "NOT_ALIVE", "182": "NOT_ALIVE", "183": "NON_PARTICIPATING", "184": "NOT_ALIVE", "185": "NON_PARTICIPATING", "186": "NON_PARTICIPATING", "187": "NON_PARTICIPATING", "188": "NOT_ALIVE", "189": "NOT_ALIVE", "19": "NOT_ALIVE", "190": "NON_PARTICIPATING", "191": "NOT_ALIVE", "192": "NOT_ALIVE", "193": "NOT_ALIVE", "194": "NOT_ALIVE", "195": "NON_PARTICIPATING", "196": "NON_PARTICIPATING", "197": "SUCCESS", "198": "NON_PARTICIPATING", "199": "NON_PARTICIPATING", "2": "NOT_ALIVE", "20": "NOT_ALIVE", "200": "NOT_ALIVE", "201": "NOT_ALIVE", "202": "NOT_ALIVE", "203": "NOT_ALIVE", "204": "NON_PARTICIPATING", "205": "NON_PARTICIPATING", "206": "NON_PARTICIPATING", "207": "NOT_ALIVE", "208": "NON_PARTICIPATING", "209": "NOT_ALIVE", "21": "NOT_ALIVE", "210": "NON_PARTICIPATING", "211": "NON_PARTICIPATING", "212": "NON_PARTICIPATING", "213": "NOT_ALIVE", "214": "NON_PARTICIPATING", "215": "NON_PARTICIPATING", "216": "NON_PARTICIPATING", "217": "NON_PARTICIPATING", "218": "NON_PARTICIPATING", "219": "NON_PARTICIPATING", "22": "NON_PARTICIPATING", "220": "NON_PARTICIPATING", "221": "NOT_ALIVE", "222": "NOT_ALIVE", "223": "NON_PARTICIPATING", "224": "NOT_ALIVE", "225": "NOT_ALIVE", "226": "NOT_ALIVE", "227": "NON_PARTICIPATING", "228": "NON_PARTICIPATING", "229": "NON_PARTICIPATING", "23": "NON_PARTICIPATING", "230": "NOT_ALIVE", "231": "NOT_ALIVE", "232": "NOT_ALIVE", "233": "NON_PARTICIPATING", "234": "NON_PARTICIPATING", "235": "NOT_ALIVE", "236": "NOT_ALIVE", "237": "SUCCESS", "238": "NOT_ALIVE", "239": "NOT_ALIVE", "24": "NOT_ALIVE", "240": "NOT_ALIVE", "241": "NOT_ALIVE", "242": "NON_PARTICIPATING", "243": "NOT_ALIVE", "244": "NON_PARTICIPATING", "245": "NON_PARTICIPATING", "246": "NON_PARTICIPATING", "247": "NON_PARTICIPATING", "248": "NOT_ALIVE", "249": "NOT_ALIVE", "25": "SUCCESS", "250": "NON_PARTICIPATING", "251": "NOT_ALIVE", "252": "NOT_ALIVE", "253": "NOT_ALIVE", "254": "NON_PARTICIPATING", "255": "NOT_ALIVE", "26": "NON_PARTICIPATING", "27": "SUCCESS", "28": "NOT_ALIVE", "29": "NON_PARTICIPATING", "3": "NON_PARTICIPATING", "30": "NON_PARTICIPATING", "31": "NON_PARTICIPATING", "32": "SUCCESS", "33": "NOT_ALIVE", "34": "NOT_ALIVE", "35": "SUCCESS", "36": "NOT_ALIVE", "37": "NOT_ALIVE", "38": "NON_PARTICIPATING", "39": "NOT_ALIVE", "4": "NON_PARTICIPATING", "40": "NON_PARTICIPATING", "41": "NOT_ALIVE", "42": "NON_PARTICIPATING", "43": "NON_PARTICIPATING", "44": "NOT_ALIVE", "45": "NOT_ALIVE", "46": "NOT_ALIVE", "47": "NON_PARTICIPATING", "48": "NON_PARTICIPATING", "49": "NON_PARTICIPATING", "5": "NON_PARTICIPATING", "50": "NON_PARTICIPATING", "51": "NON_PARTICIPATING", "52": "NOT_ALIVE", "53": "NON_PARTICIPATING", "54": "NON_PARTICIPATING", "55": "NOT_ALIVE", "56": "NON_PARTICIPATING", "57": "NOT_ALIVE", "58": "NON_PARTICIPATING", "59": "SUCCESS", "6": "NON_PARTICIPATING", "60": "NON_PARTICIPATING", "61": "NON_PARTICIPATING", "62": "NON_PARTICIPATING", "63": "NON_PARTICIPATING", "64": "NOT_ALIVE", "65": "NOT_ALIVE", "66": "NOT_ALIVE", "67": "NOT_ALIVE", "68": "NON_PARTICIPATING", "69": "NOT_ALIVE", "7": "NOT_ALIVE", "70": "NON_PARTICIPATING", "71": "NON_PARTICIPATING", "72": "NOT_ALIVE", "73": "SUCCESS", "74": "NOT_ALIVE", "75": "NOT_ALIVE", "76": "NON_PARTICIPATING", "77": "NON_PARTICIPATING", "78": "NOT_ALIVE", "79": "NOT_ALIVE", "8": "NOT_ALIVE", "80": "NON_PARTICIPATING", "81": "NON_PARTICIPATING", "82": "NON_PARTICIPATING", "83": "NON_PARTICIPATING", "84": "NON_PARTICIPATING", "85": "NON_PARTICIPATING", "86": "NOT_ALIVE", "87": "NON_PARTICIPATING", "88": "NOT_ALIVE", "89": "NOT_ALIVE", "9": "NOT_ALIVE", "90": "NON_PARTICIPATING", "91": "NOT_ALIVE", "92": "SUCCESS", "93": "NON_PARTICIPATING", "94": "SUCCESS", "95": "NOT_ALIVE", "96": "NON_PARTICIPATING", "97": "NOT_ALIVE", "98": "NON_PARTICIPATING", "99": "NOT_ALIVE" }, "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "block_list": [ 5965078, 5965092 ], "bos_token_id": 1, "eos_token_id": 2, "hidden_act": "silu", "hidden_size": 2048, "initializer_range": 0.02, "inner_step": 4, "intermediate_size": 5632, "last_allreduce_block": 5964830, "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 }