{ "_name_or_path": "distributed/llama-1b", "all_reduce_scores": { "0": "NOT_ALIVE", "1": "SUCCESS", "10": "NOT_ALIVE", "100": "NOT_ALIVE", "101": "SUCCESS", "102": "NON_PARTICIPATING", "103": "NON_PARTICIPATING", "104": "NON_PARTICIPATING", "105": "NOT_ALIVE", "106": "FAIL", "107": "SUCCESS", "108": "NON_PARTICIPATING", "109": "NON_PARTICIPATING", "11": "SUCCESS", "110": "FAIL", "111": "NOT_ALIVE", "112": "SUCCESS", "113": "NOT_ALIVE", "114": "SUCCESS", "115": "NON_PARTICIPATING", "116": "NON_PARTICIPATING", "117": "SUCCESS", "118": "NOT_ALIVE", "119": "FAIL", "12": "NOT_ALIVE", "120": "NON_PARTICIPATING", "121": "SUCCESS", "122": "NON_PARTICIPATING", "123": "NOT_ALIVE", "124": "NOT_ALIVE", "125": "NOT_ALIVE", "126": "SUCCESS", "127": "SUCCESS", "128": "SUCCESS", "129": "NOT_ALIVE", "13": "NOT_ALIVE", "130": "NOT_ALIVE", "131": "NON_PARTICIPATING", "132": "SUCCESS", "133": "SUCCESS", "134": "NOT_ALIVE", "135": "SUCCESS", "136": "SUCCESS", "137": "SUCCESS", "138": "NON_PARTICIPATING", "139": "SUCCESS", "14": "NOT_ALIVE", "140": "FAIL", "141": "NON_PARTICIPATING", "142": "NON_PARTICIPATING", "143": "NON_PARTICIPATING", "144": "FAIL", "145": "NOT_ALIVE", "146": "SUCCESS", "147": "SUCCESS", "148": "NOT_ALIVE", "149": "SUCCESS", "15": "SUCCESS", "150": "FAIL", "151": "NON_PARTICIPATING", "152": "SUCCESS", "153": "NON_PARTICIPATING", "154": "NON_PARTICIPATING", "155": "SUCCESS", "156": "NOT_ALIVE", "157": "NON_PARTICIPATING", "158": "FAIL", "159": "SUCCESS", "16": "NOT_ALIVE", "160": "NON_PARTICIPATING", "161": "FAIL", "162": "NON_PARTICIPATING", "163": "NOT_ALIVE", "164": "NON_PARTICIPATING", "165": "FAIL", "166": "NOT_ALIVE", "167": "SUCCESS", "168": "SUCCESS", "169": "SUCCESS", "17": "NOT_ALIVE", "170": "SUCCESS", "171": "SUCCESS", "172": "FAIL", "173": "SUCCESS", "174": "SUCCESS", "175": "SUCCESS", "176": "NOT_ALIVE", "177": "SUCCESS", "178": "NON_PARTICIPATING", "179": "FAIL", "18": "NOT_ALIVE", "180": "SUCCESS", "181": "SUCCESS", "182": "NOT_ALIVE", "183": "FAIL", "184": "SUCCESS", "185": "NOT_ALIVE", "186": "NON_PARTICIPATING", "187": "NOT_ALIVE", "188": "SUCCESS", "189": "SUCCESS", "19": "FAIL", "190": "NON_PARTICIPATING", "191": "FAIL", "192": "SUCCESS", "193": "NON_PARTICIPATING", "194": "NON_PARTICIPATING", "195": "NON_PARTICIPATING", "196": "SUCCESS", "197": "NON_PARTICIPATING", "198": "NON_PARTICIPATING", "199": "FAIL", "2": "NOT_ALIVE", "20": "NOT_ALIVE", "200": "SUCCESS", "201": "NOT_ALIVE", "202": "NON_PARTICIPATING", "203": "SUCCESS", "204": "SUCCESS", "205": "NON_PARTICIPATING", "206": "SUCCESS", "207": "NON_PARTICIPATING", "208": "NON_PARTICIPATING", "209": "SUCCESS", "21": "SUCCESS", "210": "NOT_ALIVE", "211": "NOT_ALIVE", "212": "NOT_ALIVE", "213": "SUCCESS", "214": "SUCCESS", "215": "FAIL", "216": "SUCCESS", "217": "SUCCESS", "218": "NOT_ALIVE", "219": "FAIL", "22": "FAIL", "220": "FAIL", "221": "NON_PARTICIPATING", "222": "NOT_ALIVE", "223": "SUCCESS", "224": "NON_PARTICIPATING", "225": "SUCCESS", "226": "NOT_ALIVE", "227": "NOT_ALIVE", "228": "NON_PARTICIPATING", "229": "NON_PARTICIPATING", "23": "NOT_ALIVE", "230": "NOT_ALIVE", "231": "NON_PARTICIPATING", "232": "FAIL", "233": "SUCCESS", "234": "SUCCESS", "235": "FAIL", "236": "SUCCESS", "237": "NOT_ALIVE", "238": "SUCCESS", "239": "FAIL", "24": "SUCCESS", "240": "NON_PARTICIPATING", "241": "FAIL", "242": "SUCCESS", "243": "NOT_ALIVE", "244": "SUCCESS", "245": "NON_PARTICIPATING", "246": "NON_PARTICIPATING", "247": "FAIL", "248": "SUCCESS", "249": "NOT_ALIVE", "25": "SUCCESS", "250": "FAIL", "251": "NOT_ALIVE", "252": "SUCCESS", "253": "NOT_ALIVE", "254": "NOT_ALIVE", "255": "NOT_ALIVE", "26": "SUCCESS", "27": "SUCCESS", "28": "SUCCESS", "29": "NON_PARTICIPATING", "3": "SUCCESS", "30": "SUCCESS", "31": "SUCCESS", "32": "FAIL", "33": "NOT_ALIVE", "34": "SUCCESS", "35": "FAIL", "36": "FAIL", "37": "NOT_ALIVE", "38": "NOT_ALIVE", "39": "NOT_ALIVE", "4": "NOT_ALIVE", "40": "NON_PARTICIPATING", "41": "SUCCESS", "42": "SUCCESS", "43": "NON_PARTICIPATING", "44": "SUCCESS", "45": "FAIL", "46": "SUCCESS", "47": "SUCCESS", "48": "NOT_ALIVE", "49": "NON_PARTICIPATING", "5": "FAIL", "50": "FAIL", "51": "NON_PARTICIPATING", "52": "SUCCESS", "53": "SUCCESS", "54": "SUCCESS", "55": "SUCCESS", "56": "NON_PARTICIPATING", "57": "NON_PARTICIPATING", "58": "FAIL", "59": "NON_PARTICIPATING", "6": "NOT_ALIVE", "60": "NON_PARTICIPATING", "61": "NON_PARTICIPATING", "62": "SUCCESS", "63": "FAIL", "64": "SUCCESS", "65": "NOT_ALIVE", "66": "NON_PARTICIPATING", "67": "NON_PARTICIPATING", "68": "SUCCESS", "69": "NON_PARTICIPATING", "7": "FAIL", "70": "SUCCESS", "71": "NON_PARTICIPATING", "72": "NON_PARTICIPATING", "73": "NON_PARTICIPATING", "74": "NON_PARTICIPATING", "75": "NON_PARTICIPATING", "76": "SUCCESS", "77": "SUCCESS", "78": "SUCCESS", "79": "FAIL", "8": "NOT_ALIVE", "80": "NON_PARTICIPATING", "81": "NON_PARTICIPATING", "82": "SUCCESS", "83": "SUCCESS", "84": "FAIL", "85": "NON_PARTICIPATING", "86": "NOT_ALIVE", "87": "SUCCESS", "88": "NON_PARTICIPATING", "89": "NON_PARTICIPATING", "9": "NOT_ALIVE", "90": "SUCCESS", "91": "FAIL", "92": "NOT_ALIVE", "93": "NON_PARTICIPATING", "94": "NOT_ALIVE", "95": "SUCCESS", "96": "FAIL", "97": "FAIL", "98": "NON_PARTICIPATING", "99": "NOT_ALIVE" }, "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "block_list": [ 6381874, 6381882, 6381890 ], "bos_token_id": 1, "eos_token_id": 2, "hidden_act": "silu", "hidden_size": 2048, "initializer_range": 0.02, "inner_step": 82, "intermediate_size": 5632, "last_allreduce_block": 6381529, "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 }