{ "_name_or_path": "distributed/llama-1b", "all_reduce_scores": { "0": "NOT_ALIVE", "1": "SUCCESS", "10": "NON_PARTICIPATING", "100": "NOT_ALIVE", "101": "NOT_ALIVE", "102": "NOT_ALIVE", "103": "NON_PARTICIPATING", "104": "NOT_ALIVE", "105": "SUCCESS", "106": "NON_PARTICIPATING", "107": "NON_PARTICIPATING", "108": "NOT_ALIVE", "109": "SUCCESS", "11": "NOT_ALIVE", "110": "NON_PARTICIPATING", "111": "NOT_ALIVE", "112": "NON_PARTICIPATING", "113": "SUCCESS", "114": "SUCCESS", "115": "NON_PARTICIPATING", "116": "NON_PARTICIPATING", "117": "NON_PARTICIPATING", "118": "NOT_ALIVE", "119": "FAIL", "12": "NOT_ALIVE", "120": "NON_PARTICIPATING", "121": "NOT_ALIVE", "122": "NON_PARTICIPATING", "123": "SUCCESS", "124": "NOT_ALIVE", "125": "NOT_ALIVE", "126": "NON_PARTICIPATING", "127": "NON_PARTICIPATING", "128": "NON_PARTICIPATING", "129": "NON_PARTICIPATING", "13": "SUCCESS", "130": "NOT_ALIVE", "131": "NOT_ALIVE", "132": "NOT_ALIVE", "133": "NON_PARTICIPATING", "134": "NOT_ALIVE", "135": "NOT_ALIVE", "136": "NOT_ALIVE", "137": "NOT_ALIVE", "138": "NON_PARTICIPATING", "139": "NON_PARTICIPATING", "14": "NON_PARTICIPATING", "140": "NOT_ALIVE", "141": "NOT_ALIVE", "142": "SUCCESS", "143": "NON_PARTICIPATING", "144": "NON_PARTICIPATING", "145": "NOT_ALIVE", "146": "SUCCESS", "147": "NON_PARTICIPATING", "148": "NON_PARTICIPATING", "149": "NOT_ALIVE", "15": "SUCCESS", "150": "SUCCESS", "151": "NOT_ALIVE", "152": "NON_PARTICIPATING", "153": "SUCCESS", "154": "SUCCESS", "155": "NOT_ALIVE", "156": "NOT_ALIVE", "157": "NON_PARTICIPATING", "158": "NOT_ALIVE", "159": "NOT_ALIVE", "16": "NOT_ALIVE", "160": "NON_PARTICIPATING", "161": "NOT_ALIVE", "162": "NON_PARTICIPATING", "163": "SUCCESS", "164": "NOT_ALIVE", "165": "NON_PARTICIPATING", "166": "NOT_ALIVE", "167": "NON_PARTICIPATING", "168": "NON_PARTICIPATING", "169": "SUCCESS", "17": "NON_PARTICIPATING", "170": "NOT_ALIVE", "171": "NON_PARTICIPATING", "172": "NOT_ALIVE", "173": "FAIL", "174": "NON_PARTICIPATING", "175": "NON_PARTICIPATING", "176": "NOT_ALIVE", "177": "SUCCESS", "178": "NOT_ALIVE", "179": "NON_PARTICIPATING", "18": "NON_PARTICIPATING", "180": "NON_PARTICIPATING", "181": "NOT_ALIVE", "182": "NON_PARTICIPATING", "183": "SUCCESS", "184": "SUCCESS", "185": "NOT_ALIVE", "186": "NOT_ALIVE", "187": "NOT_ALIVE", "188": "SUCCESS", "189": "NOT_ALIVE", "19": "NON_PARTICIPATING", "190": "SUCCESS", "191": "NOT_ALIVE", "192": "NON_PARTICIPATING", "193": "SUCCESS", "194": "NOT_ALIVE", "195": "NOT_ALIVE", "196": "SUCCESS", "197": "NON_PARTICIPATING", "198": "SUCCESS", "199": "NON_PARTICIPATING", "2": "NON_PARTICIPATING", "20": "SUCCESS", "200": "SUCCESS", "201": "NON_PARTICIPATING", "202": "NON_PARTICIPATING", "203": "NON_PARTICIPATING", "204": "NOT_ALIVE", "205": "SUCCESS", "206": "SUCCESS", "207": "NOT_ALIVE", "208": "NON_PARTICIPATING", "209": "NON_PARTICIPATING", "21": "NOT_ALIVE", "210": "NON_PARTICIPATING", "211": "NON_PARTICIPATING", "212": "SUCCESS", "213": "NON_PARTICIPATING", "214": "SUCCESS", "215": "NON_PARTICIPATING", "216": "NOT_ALIVE", "217": "NOT_ALIVE", "218": "NON_PARTICIPATING", "219": "SUCCESS", "22": "NON_PARTICIPATING", "220": "NOT_ALIVE", "221": "SUCCESS", "222": "SUCCESS", "223": "SUCCESS", "224": "NOT_ALIVE", "225": "NON_PARTICIPATING", "226": "NON_PARTICIPATING", "227": "SUCCESS", "228": "NOT_ALIVE", "229": "NON_PARTICIPATING", "23": "NON_PARTICIPATING", "230": "NOT_ALIVE", "231": "NOT_ALIVE", "232": "NON_PARTICIPATING", "233": "NON_PARTICIPATING", "234": "NON_PARTICIPATING", "235": "SUCCESS", "236": "NON_PARTICIPATING", "237": "SUCCESS", "238": "SUCCESS", "239": "NON_PARTICIPATING", "24": "NON_PARTICIPATING", "240": "NOT_ALIVE", "241": "SUCCESS", "242": "NON_PARTICIPATING", "243": "NOT_ALIVE", "244": "NOT_ALIVE", "245": "NON_PARTICIPATING", "246": "NON_PARTICIPATING", "247": "NOT_ALIVE", "248": "NON_PARTICIPATING", "249": "NON_PARTICIPATING", "25": "SUCCESS", "250": "NOT_ALIVE", "251": "NOT_ALIVE", "252": "SUCCESS", "253": "NOT_ALIVE", "254": "NON_PARTICIPATING", "255": "NON_PARTICIPATING", "26": "SUCCESS", "27": "NOT_ALIVE", "28": "SUCCESS", "29": "SUCCESS", "3": "NON_PARTICIPATING", "30": "NON_PARTICIPATING", "31": "NOT_ALIVE", "32": "NON_PARTICIPATING", "33": "NOT_ALIVE", "34": "NON_PARTICIPATING", "35": "NON_PARTICIPATING", "36": "NOT_ALIVE", "37": "NOT_ALIVE", "38": "SUCCESS", "39": "SUCCESS", "4": "SUCCESS", "40": "SUCCESS", "41": "NON_PARTICIPATING", "42": "NOT_ALIVE", "43": "NOT_ALIVE", "44": "SUCCESS", "45": "SUCCESS", "46": "SUCCESS", "47": "SUCCESS", "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": "NON_PARTICIPATING", "56": "NON_PARTICIPATING", "57": "NOT_ALIVE", "58": "NON_PARTICIPATING", "59": "NON_PARTICIPATING", "6": "NOT_ALIVE", "60": "SUCCESS", "61": "NOT_ALIVE", "62": "NON_PARTICIPATING", "63": "NON_PARTICIPATING", "64": "NOT_ALIVE", "65": "NOT_ALIVE", "66": "NON_PARTICIPATING", "67": "NON_PARTICIPATING", "68": "NOT_ALIVE", "69": "SUCCESS", "7": "SUCCESS", "70": "NOT_ALIVE", "71": "NON_PARTICIPATING", "72": "NOT_ALIVE", "73": "NON_PARTICIPATING", "74": "FAIL", "75": "NON_PARTICIPATING", "76": "NON_PARTICIPATING", "77": "NON_PARTICIPATING", "78": "NOT_ALIVE", "79": "NON_PARTICIPATING", "8": "SUCCESS", "80": "NOT_ALIVE", "81": "NOT_ALIVE", "82": "NON_PARTICIPATING", "83": "NOT_ALIVE", "84": "NOT_ALIVE", "85": "NON_PARTICIPATING", "86": "NON_PARTICIPATING", "87": "SUCCESS", "88": "NOT_ALIVE", "89": "NOT_ALIVE", "9": "SUCCESS", "90": "SUCCESS", "91": "NON_PARTICIPATING", "92": "SUCCESS", "93": "NOT_ALIVE", "94": "NON_PARTICIPATING", "95": "NOT_ALIVE", "96": "SUCCESS", "97": "NON_PARTICIPATING", "98": "NON_PARTICIPATING", "99": "NON_PARTICIPATING" }, "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "block_list": [ 6479614, 6479632 ], "bos_token_id": 1, "eos_token_id": 2, "hidden_act": "silu", "hidden_size": 2048, "initializer_range": 0.02, "inner_step": 12, "intermediate_size": 5632, "last_allreduce_block": 6479528, "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 }